20200728.md

ArXiv cs.CV --Tue, 28 Jul 2020

1.Learning Lane Graph Representations for Motion Forecasting ⬇️

We propose a motion forecasting model that exploits a novel structured map representation as well as actor-map interactions. Instead of encoding vectorized maps as raster images, we construct a lane graph from raw map data to explicitly preserve the map structure. To capture the complex topology and long range dependencies of the lane graph, we propose LaneGCN which extends graph convolutions with multiple adjacency matrices and along-lane dilation. To capture the complex interactions between actors and maps, we exploit a fusion network consisting of four types of interactions, actor-to-lane, lane-to-lane, lane-to-actor and actor-to-actor. Powered by LaneGCN and actor-map interactions, our model is able to predict accurate and realistic multi-modal trajectories. Our approach significantly outperforms the state-of-the-art on the large scale Argoverse motion forecasting benchmark.

2.Noisy Agents: Self-supervised Exploration by Predicting Auditory Events ⬇️

Humans integrate multiple sensory modalities (e.g. visual and audio) to build a causal understanding of the physical world. In this work, we propose a novel type of intrinsic motivation for Reinforcement Learning (RL) that encourages the agent to understand the causal effect of its actions through auditory event prediction. First, we allow the agent to collect a small amount of acoustic data and use K-means to discover underlying auditory event clusters. We then train a neural network to predict the auditory events and use the prediction errors as intrinsic rewards to guide RL exploration. Experimental results on Atari games show that our new intrinsic motivation significantly outperforms several state-of-the-art baselines. We further visualize our noisy agents' behavior in a physics environment and demonstrate that our newly designed intrinsic reward leads to the emergence of physical interaction behaviors (e.g. contact with objects).

3.Associative3D: Volumetric Reconstruction from Sparse Views ⬇️

This paper studies the problem of 3D volumetric reconstruction from two views of a scene with an unknown camera. While seemingly easy for humans, this problem poses many challenges for computers since it requires simultaneously reconstructing objects in the two views while also figuring out their relationship. We propose a new approach that estimates reconstructions, distributions over the camera/object and camera/camera transformations, as well as an inter-view object affinity matrix. This information is then jointly reasoned over to produce the most likely explanation of the scene. We train and test our approach on a dataset of indoor scenes, and rigorously evaluate the merits of our joint reasoning approach. Our experiments show that it is able to recover reasonable scenes from sparse views, while the problem is still challenging. Project site: this https URL

4.The Unsupervised Method of Vessel Movement Trajectory Prediction ⬇️

In real-world application scenarios, it is crucial for marine navigators and security analysts to predict vessel movement trajectories at sea based on the Automated Identification System (AIS) data in a given time span. This article presents an unsupervised method of ship movement trajectory prediction which represents the data in a three-dimensional space which consists of time difference between points, the scaled error distance between the tested and its predicted forward and backward locations, and the space-time angle. The representation feature space reduces the search scope for the next point to a collection of candidates which fit the local path prediction well, and therefore improve the accuracy. Unlike most statistical learning or deep learning methods, the proposed clustering-based trajectory reconstruction method does not require computationally expensive model training. This makes real-time reliable and accurate prediction feasible without using a training set. Our results show that the most prediction trajectories accurately consist of the true vessel paths.

5.WGANVO: Monocular Visual Odometry based on Generative Adversarial Networks ⬇️

In this work we present WGANVO, a Deep Learning based monocular Visual Odometry method. In particular, a neural network is trained to regress a pose estimate from an image pair. The training is performed using a semi-supervised approach. Unlike geometry based monocular methods, the proposed method can recover the absolute scale of the scene without neither prior knowledge nor extra information. The evaluation of the system is carried out on the well-known KITTI dataset where it is shown to work in real time and the accuracy obtained is encouraging to continue the development of Deep Learning based methods.

6.A Closer Look at Art Mediums: The MAMe Image Classification Dataset ⬇️

Art is an expression of human creativity, skill and technology. An exceptionally rich source of visual content. In the context of AI image processing systems, artworks represent one of the most challenging domains conceivable: Properly perceiving art requires attention to detail, a huge generalization capacity, and recognizing both simple and complex visual patterns. To challenge the AI community, this work introduces a novel image classification task focused on museum art mediums, the MAMe dataset. Data is gathered from three different museums, and aggregated by art experts into 29 classes of medium (i.e. materials and techniques). For each class, MAMe provides a minimum of 850 images (700 for training) of high-resolution and variable shape. The combination of volume, resolution and shape allows MAMe to fill a void in current image classification challenges, empowering research in aspects so far overseen by the research community. After reviewing the singularity of MAMe in the context of current image classification tasks, a thorough description of the task is provided, together with dataset statistics. Baseline experiments are conducted using well-known architectures, to highlight both the feasibility and complexity of the task proposed. Finally, these baselines are inspected using explainability methods and expert knowledge, to gain insight on the challenges that remain ahead.

7.Ordinary Differential Equation and Complex Matrix Exponential for Multi-resolution Image Registration ⬇️

Autograd-based software packages have recently renewed interest in image registration using homography and other geometric models by gradient descent and optimization, e.g., AirLab and DRMIME. In this work, we emphasize on using complex matrix exponential (CME) over real matrix exponential to compute transformation matrices. CME is theoretically more suitable and practically provides faster convergence as our experiments show. Further, we demonstrate that the use of an ordinary differential equation (ODE) as an optimizable dynamical system can adapt the transformation matrix more accurately to the multi-resolution Gaussian pyramid for image registration. Our experiments include four publicly available benchmark datasets, two of them 2D and the other two being 3D. Experiments demonstrate that our proposed method yields significantly better registration compared to a number of off-the-shelf, popular, state-of-the-art image registration toolboxes.

8.3D Human Shape and Pose from a Single Low-Resolution Image with Self-Supervised Learning ⬇️

3D human shape and pose estimation from monocular images has been an active area of research in computer vision, having a substantial impact on the development of new applications, from activity recognition to creating virtual avatars. Existing deep learning methods for 3D human shape and pose estimation rely on relatively high-resolution input images; however, high-resolution visual content is not always available in several practical scenarios such as video surveillance and sports broadcasting. Low-resolution images in real scenarios can vary in a wide range of sizes, and a model trained in one resolution does not typically degrade gracefully across resolutions. Two common approaches to solve the problem of low-resolution input are applying super-resolution techniques to the input images which may result in visual artifacts, or simply training one model for each resolution, which is impractical in many realistic applications. To address the above issues, this paper proposes a novel algorithm called RSC-Net, which consists of a Resolution-aware network, a Self-supervision loss, and a Contrastive learning scheme. The proposed network is able to learn the 3D body shape and pose across different resolutions with a single model. The self-supervision loss encourages scale-consistency of the output, and the contrastive learning scheme enforces scale-consistency of the deep features. We show that both these new training losses provide robustness when learning 3D shape and pose in a weakly-supervised manner. Extensive experiments demonstrate that the RSC-Net can achieve consistently better results than the state-of-the-art methods for challenging low-resolution images.

9.Solving Linear Inverse Problems Using the Prior Implicit in a Denoiser ⬇️

Prior probability models are a central component of many image processing problems, but density estimation is notoriously difficult for high-dimensional signals such as photographic images. Deep neural networks have provided state-of-the-art solutions for problems such as denoising, which implicitly rely on a prior probability model of natural images. Here, we develop a robust and general methodology for making use of this implicit prior. We rely on a little-known statistical result due to Miyasawa (1961), who showed that the least-squares solution for removing additive Gaussian noise can be written directly in terms of the gradient of the log of the noisy signal density. We use this fact to develop a stochastic coarse-to-fine gradient ascent procedure for drawing high-probability samples from the implicit prior embedded within a CNN trained to perform blind (i.e., unknown noise level) least-squares denoising. A generalization of this algorithm to constrained sampling provides a method for using the implicit prior to solve any linear inverse problem, with no additional training. We demonstrate this general form of transfer learning in multiple applications, using the same algorithm to produce high-quality solutions for deblurring, super-resolution, inpainting, and compressive sensing.

10.Message Passing Least Squares Framework and its Application to Rotation Synchronization ⬇️

We propose an efficient algorithm for solving group synchronization under high levels of corruption and noise, while we focus on rotation synchronization. We first describe our recent theoretically guaranteed message passing algorithm that estimates the corruption levels of the measured group ratios. We then propose a novel reweighted least squares method to estimate the group elements, where the weights are initialized and iteratively updated using the estimated corruption levels. We demonstrate the superior performance of our algorithm over state-of-the-art methods for rotation synchronization using both synthetic and real data.

11.Black-Box Face Recovery from Identity Features ⬇️

In this work, we present a novel algorithm based on an it-erative sampling of random Gaussian blobs for black-box face recovery,given only an output feature vector of deep face recognition systems. Weattack the state-of-the-art face recognition system (ArcFace) to test ouralgorithm. Another network with different architecture (FaceNet) is usedas an independent critic showing that the target person can be identi-fied with the reconstructed image even with no access to the attackedmodel. Furthermore, our algorithm requires a significantly less numberof queries compared to the state-of-the-art solution.

12.Towards Accuracy-Fairness Paradox: Adversarial Example-based Data Augmentation for Visual Debiasing ⬇️

Machine learning fairness concerns about the biases towards certain protected or sensitive group of people when addressing the target tasks. This paper studies the debiasing problem in the context of image classification tasks. Our data analysis on facial attribute recognition demonstrates (1) the attribution of model bias from imbalanced training data distribution and (2) the potential of adversarial examples in balancing data distribution. We are thus motivated to employ adversarial example to augment the training data for visual debiasing. Specifically, to ensure the adversarial generalization as well as cross-task transferability, we propose to couple the operations of target task classifier training, bias task classifier training, and adversarial example generation. The generated adversarial examples supplement the target task training dataset via balancing the distribution over bias variables in an online fashion. Results on simulated and real-world debiasing experiments demonstrate the effectiveness of the proposed solution in simultaneously improving model accuracy and fairness. Preliminary experiment on few-shot learning further shows the potential of adversarial attack-based pseudo sample generation as alternative solution to make up for the training data lackage.

13.SparseTrain: Exploiting Dataflow Sparsity for Efficient Convolutional Neural Networks Training ⬇️

Training Convolutional Neural Networks (CNNs) usually requires a large number of computational resources. In this paper, \textit{SparseTrain} is proposed to accelerate CNN training by fully exploiting the sparsity. It mainly involves three levels of innovations: activation gradients pruning algorithm, sparse training dataflow, and accelerator architecture. By applying a stochastic pruning algorithm on each layer, the sparsity of back-propagation gradients can be increased dramatically without degrading training accuracy and convergence rate. Moreover, to utilize both \textit{natural sparsity} (resulted from ReLU or Pooling layers) and \textit{artificial sparsity} (brought by pruning algorithm), a sparse-aware architecture is proposed for training acceleration. This architecture supports forward and back-propagation of CNN by adopting 1-Dimensional convolution dataflow. We have built %a simple compiler to map CNNs topology onto \textit{SparseTrain}, and a cycle-accurate architecture simulator to evaluate the performance and efficiency based on the synthesized design with $14nm$ FinFET technologies. Evaluation results on AlexNet/ResNet show that \textit{SparseTrain} could achieve about $2.7 \times$ speedup and $2.2 \times$ energy efficiency improvement on average compared with the original training process.

14.MADGAN: unsupervised Medical Anomaly Detection GAN using multiple adjacent brain MRI slice reconstruction ⬇️

Unsupervised learning can discover various unseen diseases, relying on large-scale unannotated medical images of healthy subjects. Towards this, unsupervised methods reconstruct a 2D/3D single medical image to detect outliers either in the learned feature space or from high reconstruction loss. However, without considering continuity between multiple adjacent slices, they cannot directly discriminate diseases composed of the accumulation of subtle anatomical anomalies, such as Alzheimer's Disease (AD). Moreover, no study has shown how unsupervised anomaly detection is associated with either disease stages, various (i.e., more than two types of) diseases, or multi-sequence Magnetic Resonance Imaging (MRI) scans. Therefore, we propose unsupervised Medical Anomaly Detection Generative Adversarial Network (MADGAN), a novel two-step method using GAN-based multiple adjacent brain MRI slice reconstruction to detect various diseases at different stages on multi-sequence structural MRI: (Reconstruction) Wasserstein loss with Gradient Penalty + 100 L1 loss-trained on 3 healthy brain axial MRI slices to reconstruct the next 3 ones-reconstructs unseen healthy/abnormal scans; (Diagnosis) Average L2 loss per scan discriminates them, comparing the ground truth/reconstructed slices. For training, we use 1,133 healthy T1-weighted (T1) and 135 healthy contrast-enhanced T1 (T1c) brain MRI scans. Our Self-Attention MADGAN can detect AD on T1 scans at a very early stage, Mild Cognitive Impairment (MCI), with Area Under the Curve (AUC) 0.727, and AD at a late stage with AUC 0.894, while detecting brain metastases on T1c scans with AUC 0.921.

15.Differentiable Manifold Reconstruction for Point Cloud Denoising ⬇️

3D point clouds are often perturbed by noise due to the inherent limitation of acquisition equipments, which obstructs downstream tasks such as surface reconstruction, rendering and so on. Previous works mostly infer the displacement of noisy points from the underlying surface, which however are not designated to recover the surface explicitly and may lead to sub-optimal denoising results. To this end, we propose to learn the underlying manifold of a noisy point cloud from differentiably subsampled points with trivial noise perturbation and their embedded neighborhood feature, aiming to capture intrinsic structures in point clouds. Specifically, we present an autoencoder-like neural network. The encoder learns both local and non-local feature representations of each point, and then samples points with low noise via an adaptive differentiable pooling operation. Afterwards, the decoder infers the underlying manifold by transforming each sampled point along with the embedded feature of its neighborhood to a local surface centered around the point. By resampling on the reconstructed manifold, we obtain a denoised point cloud. Further, we design an unsupervised training loss, so that our network can be trained in either an unsupervised or supervised fashion. Experiments show that our method significantly outperforms state-of-the-art denoising methods under both synthetic noise and real world noise. The code and data are available at this https URL

16.Reconstruction Regularized Deep Metric Learning for Multi-label Image Classification ⬇️

In this paper, we present a novel deep metric learning method to tackle the multi-label image classification problem. In order to better learn the correlations among images features, as well as labels, we attempt to explore a latent space, where images and labels are embedded via two unique deep neural networks, respectively. To capture the relationships between image features and labels, we aim to learn a \emph{two-way} deep distance metric over the embedding space from two different views, i.e., the distance between one image and its labels is not only smaller than those distances between the image and its labels' nearest neighbors, but also smaller than the distances between the labels and other images corresponding to the labels' nearest neighbors. Moreover, a reconstruction module for recovering correct labels is incorporated into the whole framework as a regularization term, such that the label embedding space is more representative. Our model can be trained in an end-to-end manner. Experimental results on publicly available image datasets corroborate the efficacy of our method compared with the state-of-the-arts.

17.A Novel adaptive optimization of Dual-Tree Complex Wavelet Transform for Medical Image Fusion ⬇️

In recent years, many research achievements are made in the medical image fusion field. Fusion is basically extraction of best of inputs and conveying it to the output. Medical Image fusion means that several of various modality image information is comprehended together to form one image to express its information. The aim of image fusion is to integrate complementary and redundant information. In this paper, a multimodal image fusion algorithm based on the dual-tree complex wavelet transform (DT-CWT) and adaptive particle swarm optimization (APSO) is proposed. Fusion is achieved through the formation of a fused pyramid using the DTCWT coefficients from the decomposed pyramids of the source images. The coefficients are fused by the weighted average method based on pixels, and the weights are estimated by the APSO to gain optimal fused images. The fused image is obtained through conventional inverse dual-tree complex wavelet transform reconstruction process. Experiment results show that the proposed method based on adaptive particle swarm optimization algorithm is remarkably better than the method based on particle swarm optimization. The resulting fused images are compared visually and through benchmarks such as Entropy (E), Peak Signal to Noise Ratio, (PSNR), Root Mean Square Error (RMSE), Standard deviation (SD) and Structure Similarity Index Metric (SSIM) computations.

18.The Effect of Wearing a Mask on Face Recognition Performance: an Exploratory Study ⬇️

Face recognition has become essential in our daily lives as a convenient and contactless method of accurate identity verification. Process such as identity verification at automatic border control gates or the secure login to electronic devices are increasingly dependant on such technologies. The recent COVID-19 pandemic have increased the value of hygienic and contactless identity verification. However, the pandemic led to the wide use of face masks, essential to keep the pandemic under control. The effect of wearing a mask on face recognition in a collaborative environment is currently sensitive yet understudied issue. We address that by presenting a specifically collected database containing three session, each with three different capture instructions, to simulate realistic use cases. We further study the effect of masked face probes on the behaviour of three top-performing face recognition systems, two academic solutions and one commercial off-the-shelf (COTS) system.

19.Identity-Guided Human Semantic Parsing for Person Re-Identification ⬇️

Existing alignment-based methods have to employ the pretrained human parsing models to achieve the pixel-level alignment, and cannot identify the personal belongings (e.g., backpacks and reticule) which are crucial to person re-ID. In this paper, we propose the identity-guided human semantic parsing approach (ISP) to locate both the human body parts and personal belongings at pixel-level for aligned person re-ID only with person identity labels. We design the cascaded clustering on feature maps to generate the pseudo-labels of human parts. Specifically, for the pixels of all images of a person, we first group them to foreground or background and then group the foreground pixels to human parts. The cluster assignments are subsequently used as pseudo-labels of human parts to supervise the part estimation and ISP iteratively learns the feature maps and groups them. Finally, local features of both human body parts and personal belongings are obtained according to the selflearned part estimation, and only features of visible parts are utilized for the retrieval. Extensive experiments on three widely used datasets validate the superiority of ISP over lots of state-of-the-art methods. Our code is available at this https URL.

20.Two-Level Residual Distillation based Triple Network for Incremental Object Detection ⬇️

Modern object detection methods based on convolutional neural network suffer from severe catastrophic forgetting in learning new classes without original data. Due to time consumption, storage burden and privacy of old data, it is inadvisable to train the model from scratch with both old and new data when new object classes emerge after the model trained. In this paper, we propose a novel incremental object detector based on Faster R-CNN to continuously learn from new object classes without using old data. It is a triple network where an old model and a residual model as assistants for helping the incremental model learning on new classes without forgetting the previous learned knowledge. To better maintain the discrimination of features between old and new classes, the residual model is jointly trained on new classes in the incremental learning procedure. In addition, a corresponding distillation scheme is designed to guide the training process, which consists of a two-level residual distillation loss and a joint classification distillation loss. Extensive experiments on VOC2007 and COCO are conducted, and the results demonstrate that the proposed method can effectively learn to incrementally detect objects of new classes, and the problem of catastrophic forgetting is mitigated in this context.

21.Contraction Mapping of Feature Norms for Classifier Learning on the Data with Different Quality ⬇️

The popular softmax loss and its recent extensions have achieved great success in the deep learning-based image clas-sification. However, the data for training image classifiers usually has different quality. Ignoring such problem, the cor-rect classification of low quality data is hard to be solved. In this paper, we discover the positive correlation between the feature norm of an image and its quality through careful ex-periments on various applications and various deep neural networks. Based on this finding, we propose a contraction mapping function to compress the range of feature norms of training images according to their quality and embed this con-traction mapping function into softmax loss or its extensions to produce novel learning objectives. The experiments on var-ious classification applications, including handwritten digit recognition, lung nodule classification, face verification and face recognition, demonstrate that the proposed approach is promising to effectively deal with the problem of learning on the data with different quality and leads to the significant and stable improvements in the classification accuracy.

22.YOLOpeds: Efficient Real-Time Single-Shot Pedestrian Detection for Smart Camera Applications ⬇️

Deep Learning-based object detectors can enhance the capabilities of smart camera systems in a wide spectrum of machine vision applications including video surveillance, autonomous driving, robots and drones, smart factory, and health monitoring. Pedestrian detection plays a key role in all these applications and deep learning can be used to construct accurate state-of-the-art detectors. However, such complex paradigms do not scale easily and are not traditionally implemented in resource-constrained smart cameras for on-device processing which offers significant advantages in situations when real-time monitoring and robustness are vital. Efficient neural networks can not only enable mobile applications and on-device experiences but can also be a key enabler of privacy and security allowing a user to gain the benefits of neural networks without needing to send their data to the server to be evaluated. This work addresses the challenge of achieving a good trade-off between accuracy and speed for efficient deployment of deep-learning-based pedestrian detection in smart camera applications. A computationally efficient architecture is introduced based on separable convolutions and proposes integrating dense connections across layers and multi-scale feature fusion to improve representational capacity while decreasing the number of parameters and operations. In particular, the contributions of this work are the following: 1) An efficient backbone combining multi-scale feature operations, 2) a more elaborate loss function for improved localization, 3) an anchor-less approach for detection, The proposed approach called YOLOpeds is evaluated using the PETS2009 surveillance dataset on 320x320 images. Overall, YOLOpeds provides real-time sustained operation of over 30 frames per second with detection rates in the range of 86% outperforming existing deep learning models.

23.Ladybird: Quasi-Monte Carlo Sampling for Deep Implicit Field Based 3D Reconstruction with Symmetry ⬇️

Deep implicit field regression methods are effective for 3D reconstruction from single-view images. However, the impact of different sampling patterns on the reconstruction quality is not well-understood. In this work, we first study the effect of point set discrepancy on the network training. Based on Farthest Point Sampling algorithm, we propose a sampling scheme that theoretically encourages better generalization performance, and results in fast convergence for SGD-based optimization algorithms. Secondly, based on the reflective symmetry of an object, we propose a feature fusion method that alleviates issues due to self-occlusions which makes it difficult to utilize local image features. Our proposed system Ladybird is able to create high quality 3D object reconstructions from a single input image. We evaluate Ladybird on a large scale 3D dataset (ShapeNet) demonstrating highly competitive results in terms of Chamfer distance, Earth Mover's distance and Intersection Over Union (IoU).

24.NOH-NMS: Improving Pedestrian Detection by Nearby Objects Hallucination ⬇️

Greedy-NMS inherently raises a dilemma, where a lower NMS threshold will potentially lead to a lower recall rate and a higher threshold introduces more false positives. This problem is more severe in pedestrian detection because the instance density varies more intensively. However, previous works on NMS don't consider or vaguely consider the factor of the existent of nearby pedestrians. Thus, we propose Nearby Objects Hallucinator (NOH), which pinpoints the objects nearby each proposal with a Gaussian distribution, together with NOH-NMS, which dynamically eases the suppression for the space that might contain other objects with a high likelihood. Compared to Greedy-NMS, our method, as the state-of-the-art, improves by $3.9%$ AP, $5.1%$ Recall, and $0.8%$ $\text{MR}^{-2}$ on CrowdHuman to $89.0%$ AP and $92.9%$ Recall, and $43.9%$ $\text{MR}^{-2}$ respectively.

25.Decomposed Generation Networks with Structure Prediction for Recipe Generation from Food Images ⬇️

Recipe generation from food images and ingredients is a challenging task, which requires the interpretation of the information from another modality. Different from the image captioning task, where the captions usually have one sentence, cooking instructions contain multiple sentences and have obvious structures. To help the model capture the recipe structure and avoid missing some cooking details, we propose a novel framework: Decomposed Generation Networks (DGN) with structure prediction, to get more structured and complete recipe generation outputs. To be specific, we split each cooking instruction into several phases, and assign different sub-generators to each phase. Our approach includes two novel ideas: (i) learning the recipe structures with the global structure prediction component and (ii) producing recipe phases in the sub-generator output component based on the predicted structure. Extensive experiments on the challenging large-scale Recipe1M dataset validate the effectiveness of our proposed model DGN, which improves the performance over the state-of-the-art results.

26.Part-Aware Data Augmentation for 3D Object Detection in Point Cloud ⬇️

Data augmentation has greatly contributed to improving the performance in image recognition tasks, and a lot of related studies have been conducted. However, data augmentation on 3D point cloud data has not been much explored. 3D label has more sophisticated and rich structural information than the 2D label, so it enables more diverse and effective data augmentation. In this paper, we propose part-aware data augmentation (PA-AUG) that can better utilize rich information of 3D label to enhance the performance of 3D object detectors. PA-AUG divides objects into partitions and stochastically applies five novel augmentation methods to each local region. It is compatible with existing point cloud data augmentation methods and can be used universally regardless of the detector's architecture. PA-AUG has improved the performance of state-of-the-art 3D object detector for all classes of the KITTI dataset and has the equivalent effect of increasing the train data by about 2.5$\times$. We also show that PA-AUG not only increases performance for a given dataset but also is robust to corrupted data. CODE WILL BE AVAILABLE.

27.Feature visualization of Raman spectrum analysis with deep convolutional neural network ⬇️

We demonstrate a recognition and feature visualization method that uses a deep convolutional neural network for Raman spectrum analysis. The visualization is achieved by calculating important regions in the spectra from weights in pooling and fully-connected layers. The method is first examined for simple Lorentzian spectra, then applied to the spectra of pharmaceutical compounds and numerically mixed amino acids. We investigate the effects of the size and number of convolution filters on the extracted regions for Raman-peak signals using the Lorentzian spectra. It is confirmed that the Raman peak contributes to the recognition by visualizing the extracted features. A near-zero weight value is obtained at the background level region, which appears to be used for baseline correction. Common component extraction is confirmed by an evaluation of numerically mixed amino acid spectra. High weight values at the common peaks and negative values at the distinctive peaks appear, even though the model is given one-hot vectors as the training labels (without a mix ratio). This proposed method is potentially suitable for applications such as the validation of trained models, ensuring the reliability of common component extraction from compound samples for spectral analysis.

28.Self-Prediction for Joint Instance and Semantic Segmentation of Point Clouds ⬇️

We develop a novel learning scheme named Self-Prediction for 3D instance and semantic segmentation of point clouds. Distinct from most existing methods that focus on designing convolutional operators, our method designs a new learning scheme to enhance point relation exploring for better segmentation. More specifically, we divide a point cloud sample into two subsets and construct a complete graph based on their representations. Then we use label propagation algorithm to predict labels of one subset when given labels of the other subset. By training with this Self-Prediction task, the backbone network is constrained to fully explore relational context/geometric/shape information and learn more discriminative features for segmentation. Moreover, a general associated framework equipped with our Self-Prediction scheme is designed for enhancing instance and semantic segmentation simultaneously, where instance and semantic representations are combined to perform Self-Prediction. Through this way, instance and semantic segmentation are collaborated and mutually reinforced. Significant performance improvements on instance and semantic segmentation compared with baseline are achieved on S3DIS and ShapeNet. Our method achieves state-of-the-art instance segmentation results on S3DIS and comparable semantic segmentation results compared with state-of-the-arts on S3DIS and ShapeNet when we only take PointNet++ as the backbone network.

29.Few-shot Knowledge Transfer for Fine-grained Cartoon Face Generation ⬇️

In this paper, we are interested in generating fine-grained cartoon faces for various groups. We assume that one of these groups consists of sufficient training data while the others only contain few samples. Although the cartoon faces of these groups share similar style, the appearances in various groups could still have some specific characteristics, which makes them differ from each other. A major challenge of this task is how to transfer knowledge among groups and learn group-specific characteristics with only few samples. In order to solve this problem, we propose a two-stage training process. First, a basic translation model for the basic group (which consists of sufficient data) is trained. Then, given new samples of other groups, we extend the basic model by creating group-specific branches for each new group. Group-specific branches are updated directly to capture specific appearances for each group while the remaining group-shared parameters are updated indirectly to maintain the distribution of intermediate feature space. In this manner, our approach is capable to generate high-quality cartoon faces for various groups.

30.Rethinking Generative Zero-Shot Learning: An Ensemble Learning Perspective for Recognising Visual Patches ⬇️

Zero-shot learning (ZSL) is commonly used to address the very pervasive problem of predicting unseen classes in fine-grained image classification and other tasks. One family of solutions is to learn synthesised unseen visual samples produced by generative models from auxiliary semantic information, such as natural language descriptions. However, for most of these models, performance suffers from noise in the form of irrelevant image backgrounds. Further, most methods do not allocate a calculated weight to each semantic patch. Yet, in the real world, the discriminative power of features can be quantified and directly leveraged to improve accuracy and reduce computational complexity. To address these issues, we propose a novel framework called multi-patch generative adversarial nets (MPGAN) that synthesises local patch features and labels unseen classes with a novel weighted voting strategy. The process begins by generating discriminative visual features from noisy text descriptions for a set of predefined local patches using multiple specialist generative models. The features synthesised from each patch for unseen classes are then used to construct an ensemble of diverse supervised classifiers, each corresponding to one local patch. A voting strategy averages the probability distributions output from the classifiers and, given that some patches are more discriminative than others, a discrimination-based attention mechanism helps to weight each patch accordingly. Extensive experiments show that MPGAN has significantly greater accuracy than state-of-the-art methods.

31.Split Computing for Complex Object Detectors: Challenges and Preliminary Results ⬇️

Following the trends of mobile and edge computing for DNN models, an intermediate option, split computing, has been attracting attentions from the research community. Previous studies empirically showed that while mobile and edge computing often would be the best options in terms of total inference time, there are some scenarios where split computing methods can achieve shorter inference time. All the proposed split computing approaches, however, focus on image classification tasks, and most are assessed with small datasets that are far from the practical scenarios. In this paper, we discuss the challenges in developing split computing methods for powerful R-CNN object detectors trained on a large dataset, COCO 2017. We extensively analyze the object detectors in terms of layer-wise tensor size and model size, and show that naive split computing methods would not reduce inference time. To the best of our knowledge, this is the first study to inject small bottlenecks to such object detectors and unveil the potential of a split computing approach. The source code and trained models' weights used in this study are available at this https URL .

32.K-Shot Contrastive Learning of Visual Features with Multiple Instance Augmentations ⬇️

In this paper, we propose the $K$-Shot Contrastive Learning (KSCL) of visual features by applying multiple augmentations to investigate the sample variations within individual instances. It aims to combine the advantages of inter-instance discrimination by learning discriminative features to distinguish between different instances, as well as intra-instance variations by matching queries against the variants of augmented samples over instances. Particularly, for each instance, it constructs an instance subspace to model the configuration of how the significant factors of variations in $K$-shot augmentations can be combined to form the variants of augmentations. Given a query, the most relevant variant of instances is then retrieved by projecting the query onto their subspaces to predict the positive instance class. This generalizes the existing contrastive learning that can be viewed as a special one-shot case. An eigenvalue decomposition is performed to configure instance subspaces, and the embedding network can be trained end-to-end through the differentiable subspace configuration. Experiment results demonstrate the proposed $K$-shot contrastive learning achieves superior performances to the state-of-the-art unsupervised methods.

33.Reconstructing NBA Players ⬇️

Great progress has been made in 3D body pose and shape estimation from a single photo. Yet, state-of-the-art results still suffer from errors due to challenging body poses, modeling clothing, and self occlusions. The domain of basketball games is particularly challenging, as it exhibits all of these challenges. In this paper, we introduce a new approach for reconstruction of basketball players that outperforms the state-of-the-art. Key to our approach is a new method for creating poseable, skinned models of NBA players, and a large database of meshes (derived from the NBA2K19 video game), that we are releasing to the research community. Based on these models, we introduce a new method that takes as input a single photo of a clothed player in any basketball pose and outputs a high resolution mesh and 3D pose for that player. We demonstrate substantial improvement over state-of-the-art, single-image methods for body shape reconstruction.

34.Research Progress of Convolutional Neural Network and its Application in Object Detection ⬇️

With the improvement of computer performance and the increase of data volume, the object detection based on convolutional neural network (CNN) has become the main algorithm for object detection. This paper summarizes the research progress of convolutional neural networks and their applications in object detection, and focuses on analyzing and discussing a specific idea and method of applying convolutional neural networks for object detection, pointing out the current deficiencies and future development direction.

35.Representation Learning with Video Deep InfoMax ⬇️

Self-supervised learning has made unsupervised pretraining relevant again for difficult computer vision tasks. The most effective self-supervised methods involve prediction tasks based on features extracted from diverse views of the data. DeepInfoMax (DIM) is a self-supervised method which leverages the internal structure of deep networks to construct such views, forming prediction tasks between local features which depend on small patches in an image and global features which depend on the whole image. In this paper, we extend DIM to the video domain by leveraging similar structure in spatio-temporal networks, producing a method we call Video Deep InfoMax(VDIM). We find that drawing views from both natural-rate sequences and temporally-downsampled sequences yields results on Kinetics-pretrained action recognition tasks which match or outperform prior state-of-the-art methods that use more costly large-time-scale transformer models. We also examine the effects of data augmentation and fine-tuning methods, accomplishingSoTA by a large margin when training only on the UCF-101 dataset.

36.Learning Task-oriented Disentangled Representations for Unsupervised Domain Adaptation ⬇️

Unsupervised domain adaptation (UDA) aims to address the domain-shift problem between a labeled source domain and an unlabeled target domain. Many efforts have been made to address the mismatch between the distributions of training and testing data, but unfortunately, they ignore the task-oriented information across domains and are inflexible to perform well in complicated open-set scenarios. Many efforts have been made to eliminate the mismatch between the distributions of training and testing data by learning domain-invariant representations. However, the learned representations are usually not task-oriented, i.e., being class-discriminative and domain-transferable simultaneously. This drawback limits the flexibility of UDA in complicated open-set tasks where no labels are shared between domains. In this paper, we break the concept of task-orientation into task-relevance and task-irrelevance, and propose a dynamic task-oriented disentangling network (DTDN) to learn disentangled representations in an end-to-end fashion for UDA. The dynamic disentangling network effectively disentangles data representations into two components: the task-relevant ones embedding critical information associated with the task across domains, and the task-irrelevant ones with the remaining non-transferable or disturbing information. These two components are regularized by a group of task-specific objective functions across domains. Such regularization explicitly encourages disentangling and avoids the use of generative models or decoders. Experiments in complicated, open-set scenarios (retrieval tasks) and empirical benchmarks (classification tasks) demonstrate that the proposed method captures rich disentangled information and achieves superior performance.

37.REXUP: I REason, I EXtract, I UPdate with Structured Compositional Reasoning for Visual Question Answering ⬇️

Visual question answering (VQA) is a challenging multi-modal task that requires not only the semantic understanding of both images and questions, but also the sound perception of a step-by-step reasoning process that would lead to the correct answer. So far, most successful attempts in VQA have been focused on only one aspect, either the interaction of visual pixel features of images and word features of questions, or the reasoning process of answering the question in an image with simple objects. In this paper, we propose a deep reasoning VQA model with explicit visual structure-aware textual information, and it works well in capturing step-by-step reasoning process and detecting a complex object-relationship in photo-realistic images. REXUP network consists of two branches, image object-oriented and scene graph oriented, which jointly works with super-diagonal fusion compositional attention network. We quantitatively and qualitatively evaluate REXUP on the GQA dataset and conduct extensive ablation studies to explore the reasons behind REXUP's effectiveness. Our best model significantly outperforms the precious state-of-the-art, which delivers 92.7% on the validation set and 73.1% on the test-dev set.

38.Point-to-set distance functions for weakly supervised segmentation ⬇️

When pixel-level masks or partial annotations are not available for training neural networks for semantic segmentation, it is possible to use higher-level information in the form of bounding boxes, or image tags. In the imaging sciences, many applications do not have an object-background structure and bounding boxes are not available. Any available annotation typically comes from ground truth or domain experts. A direct way to train without masks is using prior knowledge on the size of objects/classes in the segmentation. We present a new algorithm to include such information via constraints on the network output, implemented via projection-based point-to-set distance functions. This type of distance functions always has the same functional form of the derivative, and avoids the need to adapt penalty functions to different constraints, as well as issues related to constraining properties typically associated with non-differentiable functions. Whereas object size information is known to enable object segmentation from bounding boxes from datasets with many general and medical images, we show that the applications extend to the imaging sciences where data represents indirect measurements, even in the case of single examples. We illustrate the capabilities in case of a) one or more classes do not have any annotation; b) there is no annotation at all; c) there are bounding boxes. We use data for hyperspectral time-lapse imaging, object segmentation in corrupted images, and sub-surface aquifer mapping from airborne-geophysical remote-sensing data. The examples verify that the developed methodology alleviates difficulties with annotating non-visual imagery for a range of experimental settings.

39.OASIS: A Large-Scale Dataset for Single Image 3D in the Wild ⬇️

Single-view 3D is the task of recovering 3D properties such as depth and surface normals from a single image. We hypothesize that a major obstacle to single-image 3D is data. We address this issue by presenting Open Annotations of Single Image Surfaces (OASIS), a dataset for single-image 3D in the wild consisting of annotations of detailed 3D geometry for 140,000 images. We train and evaluate leading models on a variety of single-image 3D tasks. We expect OASIS to be a useful resource for 3D vision research. Project site: this https URL.

40.Learning and aggregating deep local descriptors for instance-level recognition ⬇️

We propose an efficient method to learn deep local descriptors for instance-level recognition. The training only requires examples of positive and negative image pairs and is performed as metric learning of sum-pooled global image descriptors. At inference, the local descriptors are provided by the activations of internal components of the network. We demonstrate why such an approach learns local descriptors that work well for image similarity estimation with classical efficient match kernel methods. The experimental validation studies the trade-off between performance and memory requirements of the state-of-the-art image search approach based on match kernels. Compared to existing local descriptors, the proposed ones perform better in two instance-level recognition tasks and keep memory requirements lower. We experimentally show that global descriptors are not effective enough at large scale and that local descriptors are essential. We achieve state-of-the-art performance, in some cases even with a backbone network as small as ResNet18.

41.Deep Photometric Stereo for Non-Lambertian Surfaces ⬇️

This paper addresses the problem of photometric stereo, in both calibrated and uncalibrated scenarios, for non-Lambertian surfaces based on deep learning. We first introduce a fully convolutional deep network for calibrated photometric stereo, which we call PS-FCN. Unlike traditional approaches that adopt simplified reflectance models to make the problem tractable, our method directly learns the mapping from reflectance observations to surface normal, and is able to handle surfaces with general and unknown isotropic reflectance. At test time, PS-FCN takes an arbitrary number of images and their associated light directions as input and predicts a surface normal map of the scene in a fast feed-forward pass. To deal with the uncalibrated scenario where light directions are unknown, we introduce a new convolutional network, named LCNet, to estimate light directions from input images. The estimated light directions and the input images are then fed to PS-FCN to determine the surface normals. Our method does not require a pre-defined set of light directions and can handle multiple images in an order-agnostic manner. Thorough evaluation of our approach on both synthetic and real datasets shows that it outperforms state-of-the-art methods in both calibrated and uncalibrated scenarios.

42.Challenge-Aware RGBT Tracking ⬇️

RGB and thermal source data suffer from both shared and specific challenges, and how to explore and exploit them plays a critical role to represent the target appearance in RGBT tracking. In this paper, we propose a novel challenge-aware neural network to handle the modality-shared challenges (e.g., fast motion, scale variation and occlusion) and the modality-specific ones (e.g., illumination variation and thermal crossover) for RGBT tracking. In particular, we design several parameter-shared branches in each layer to model the target appearance under the modality-shared challenges, and several parameterindependent branches under the modality-specific ones. Based on the observation that the modality-specific cues of different modalities usually contains the complementary advantages, we propose a guidance module to transfer discriminative features from one modality to another one, which could enhance the discriminative ability of some weak modality. Moreover, all branches are aggregated together in an adaptive manner and parallel embedded in the backbone network to efficiently form more discriminative target representations. These challenge-aware branches are able to model the target appearance under certain challenges so that the target representations can be learnt by a few parameters even in the situation of insufficient training data. From the experimental results we will show that our method operates at a real-time speed while performing well against the state-of-the-art methods on three benchmark datasets.

43.Virtual Multi-view Fusion for 3D Semantic Segmentation ⬇️

Semantic segmentation of 3D meshes is an important problem for 3D scene understanding. In this paper we revisit the classic multiview representation of 3D meshes and study several techniques that make them effective for 3D semantic segmentation of meshes. Given a 3D mesh reconstructed from RGBD sensors, our method effectively chooses different virtual views of the 3D mesh and renders multiple 2D channels for training an effective 2D semantic segmentation model. Features from multiple per view predictions are finally fused on 3D mesh vertices to predict mesh semantic segmentation labels. Using the large scale indoor 3D semantic segmentation benchmark of ScanNet, we show that our virtual views enable more effective training of 2D semantic segmentation networks than previous multiview approaches. When the 2D per pixel predictions are aggregated on 3D surfaces, our virtual multiview fusion method is able to achieve significantly better 3D semantic segmentation results compared to all prior multiview approaches and competitive with recent 3D convolution approaches.

44.Contrastive Visual-Linguistic Pretraining ⬇️

Several multi-modality representation learning approaches such as LXMERT and ViLBERT have been proposed recently. Such approaches can achieve superior performance due to the high-level semantic information captured during large-scale multimodal pretraining. However, as ViLBERT and LXMERT adopt visual region regression and classification loss, they often suffer from domain gap and noisy label problems, based on the visual features having been pretrained on the Visual Genome dataset. To overcome these issues, we propose unbiased Contrastive Visual-Linguistic Pretraining (CVLP), which constructs a visual self-supervised loss built upon contrastive learning. We evaluate CVLP on several down-stream tasks, including VQA, GQA and NLVR2 to validate the superiority of contrastive learning on multi-modality representation learning. Our code is available at: this https URL.

45.GSNet: Joint Vehicle Pose and Shape Reconstruction with Geometrical and Scene-aware Supervision ⬇️

We present a novel end-to-end framework named as GSNet (Geometric and Scene-aware Network), which jointly estimates 6DoF poses and reconstructs detailed 3D car shapes from single urban street view. GSNet utilizes a unique four-way feature extraction and fusion scheme and directly regresses 6DoF poses and shapes in a single forward pass. Extensive experiments show that our diverse feature extraction and fusion scheme can greatly improve model performance. Based on a divide-and-conquer 3D shape representation strategy, GSNet reconstructs 3D vehicle shape with great detail (1352 vertices and 2700 faces). This dense mesh representation further leads us to consider geometrical consistency and scene context, and inspires a new multi-objective loss function to regularize network training, which in turn improves the accuracy of 6D pose estimation and validates the merit of jointly performing both tasks. We evaluate GSNet on the largest multi-task ApolloCar3D benchmark and achieve state-of-the-art performance both quantitatively and qualitatively. Project page is available at this https URL.

46.Towards End-to-end Video-based Eye-Tracking ⬇️

Estimating eye-gaze from images alone is a challenging task, in large parts due to un-observable person-specific factors. Achieving high accuracy typically requires labeled data from test users which may not be attainable in real applications. We observe that there exists a strong relationship between what users are looking at and the appearance of the user's eyes. In response to this understanding, we propose a novel dataset and accompanying method which aims to explicitly learn these semantic and temporal relationships. Our video dataset consists of time-synchronized screen recordings, user-facing camera views, and eye gaze data, which allows for new benchmarks in temporal gaze tracking as well as label-free refinement of gaze. Importantly, we demonstrate that the fusion of information from visual stimuli as well as eye images can lead towards achieving performance similar to literature-reported figures acquired through supervised personalization. Our final method yields significant performance improvements on our proposed EVE dataset, with up to a 28 percent improvement in Point-of-Gaze estimates (resulting in 2.49 degrees in angular error), paving the path towards high-accuracy screen-based eye tracking purely from webcam sensors. The dataset and reference source code are available at this https URL

47.SADet: Learning An Efficient and Accurate Pedestrian Detector ⬇️

Although the anchor-based detectors have taken a big step forward in pedestrian detection, the overall performance of algorithm still needs further improvement for practical applications, \emph{e.g.}, a good trade-off between the accuracy and efficiency. To this end, this paper proposes a series of systematic optimization strategies for the detection pipeline of one-stage detector, forming a single shot anchor-based detector (SADet) for efficient and accurate pedestrian detection, which includes three main improvements. Firstly, we optimize the sample generation process by assigning soft tags to the outlier samples to generate semi-positive samples with continuous tag value between $0$ and $1$, which not only produces more valid samples, but also strengthens the robustness of the model. Secondly, a novel Center-$IoU$ loss is applied as a new regression loss for bounding box regression, which not only retains the good characteristics of IoU loss, but also solves some defects of it. Thirdly, we also design Cosine-NMS for the postprocess of predicted bounding boxes, and further propose adaptive anchor matching to enable the model to adaptively match the anchor boxes to full or visible bounding boxes according to the degree of occlusion, making the NMS and anchor matching algorithms more suitable for occluded pedestrian detection. Though structurally simple, it presents state-of-the-art result and real-time speed of $20$ FPS for VGA-resolution images ($640 \times 480$) on challenging pedestrian detection benchmarks, i.e., CityPersons, Caltech, and human detection benchmark CrowdHuman, leading to a new attractive pedestrian detector.

48.Detection and Annotation of Plant Organs from Digitized Herbarium Scans using Deep Learning ⬇️

As herbarium specimens are increasingly becoming digitized and accessible in online repositories, advanced computer vision techniques are being used to extract information from them. The presence of certain plant organs on herbarium sheets is useful information in various scientific contexts and automatic recognition of these organs will help mobilize such information. In our study we use deep learning to detect plant organs on digitized herbarium specimens with Faster R-CNN. For our experiment we manually annotated hundreds of herbarium scans with thousands of bounding boxes for six types of plant organs and used them for training and evaluating the plant organ detection model. The model worked particularly well on leaves and stems, while flowers were also present in large numbers in the sheets, but not equally well recognized.

49.Towards Purely Unsupervised Disentanglement of Appearance and Shape for Person Images Generation ⬇️

There have been a fairly of research interests in exploring the disentanglement of appearance and shape from human images. Most existing endeavours pursuit this goal by either using training images with annotations or regulating the training process with external clues such as human skeleton, body segmentation or cloth patches etc. In this paper, we aim to address this challenge in a more unsupervised manner---we do not require any annotation nor any external task-specific clues. To this end, we formulate an encoder-decoder-like network to extract both the shape and appearance features from input images at the same time, and train the parameters by three losses: feature adversarial loss, color consistency loss and reconstruction loss. The feature adversarial loss mainly impose little to none mutual information between the extracted shape and appearance features, while the color consistency loss is to encourage the invariance of person appearance conditioned on different shapes. More importantly, our unsupervised\footnote{Unsupervised learning has many interpretations in different tasks. To be clear, in this paper, we refer unsupervised learning as learning without task-specific human annotations, pairs or any form of weak supervision.} framework utilizes learned shape features as masks which are applied to the input itself in order to obtain clean appearance features. Without using fixed input human skeleton, our network better preserves the conditional human posture while requiring less supervision. Experimental results on DeepFashion and Market1501 demonstrate that the proposed method achieves clean disentanglement and is able to synthesis novel images of comparable quality with state-of-the-art weakly-supervised or even supervised methods.

50.U2-ONet: A Two-level Nested Octave U-structure with Multiscale Attention Mechanism for Moving Instances Segmentation ⬇️

Most scenes in practical applications are dynamic scenes containing moving objects, so segmenting accurately moving objects is crucial for many computer vision applications. In order to efficiently segment out all moving objects in the scene, regardless of whether the object has a predefined semantic label, we propose a two-level nested Octave U-structure network with a multiscale attention mechanism called U2-ONet. Each stage of U2-ONet is filled with our newly designed Octave ReSidual U-block (ORSU) to enhance the ability to obtain more context information at different scales while reducing spatial redundancy of feature maps. In order to efficiently train our multi-scale deep network, we introduce a hierarchical training supervision strategy that calculates the loss at each level while adding a knowledge matching loss to keep the optimization consistency. Experimental results show that our method achieves state-of-the-art performance in several general moving objects segmentation datasets.

51.SMART: Simultaneous Multi-Agent Recurrent Trajectory Prediction ⬇️

We propose advances that address two key challenges in future trajectory prediction: (i) multimodality in both training data and predictions and (ii) constant time inference regardless of number of agents. Existing trajectory predictions are fundamentally limited by lack of diversity in training data, which is difficult to acquire with sufficient coverage of possible modes. Our first contribution is an automatic method to simulate diverse trajectories in the top-view. It uses pre-existing datasets and maps as initialization, mines existing trajectories to represent realistic driving behaviors and uses a multi-agent vehicle dynamics simulator to generate diverse new trajectories that cover various modes and are consistent with scene layout constraints. Our second contribution is a novel method that generates diverse predictions while accounting for scene semantics and multi-agent interactions, with constant-time inference independent of the number of agents. We propose a convLSTM with novel state pooling operations and losses to predict scene-consistent states of multiple agents in a single forward pass, along with a CVAE for diversity. We validate our proposed multi-agent trajectory prediction approach by training and testing on the proposed simulated dataset and existing real datasets of traffic scenes. In both cases, our approach outperforms SOTA methods by a large margin, highlighting the benefits of both our diverse dataset simulation and constant-time diverse trajectory prediction methods.

52.Approaches of large-scale images recognition with more than 50,000 categoris ⬇️

Though current CV models have been able to achieve high levels of accuracy on small-scale images classification dataset with hundreds or thousands of categories, many models become infeasible in computational or space consumption when it comes to large-scale dataset with more than 50,000 categories. In this paper, we provide a viable solution for classifying large-scale species datasets using traditional CV techniques such as.features extraction and processing, BOVW(Bag of Visual Words) and some statistical learning technics like Mini-Batch K-Means,SVM which are used in our works. And then mixed with a neural network model. When applying these techniques, we have done some optimization in time and memory consumption, so that it can be feasible for large-scale dataset. And we also use some technics to reduce the impact of mislabeling data. We use a dataset with more than 50, 000 categories, and all operations are done on common computer with l 6GB RAM and a CPU of 3. OGHz. Our contributions are: 1) analysis what problems may meet in the training processes, and presents several feasible ways to solve these problems. 2) Make traditional CV models combined with neural network models provide some feasible scenarios for training large-scale classified datasets within the constraints of time and spatial resources.

53.A Dual Iterative Refinement Method for Non-rigid Shape Matching ⬇️

In this work, a simple and efficient dual iterative refinement (DIR) method is proposed for dense correspondence between two nearly isometric shapes. The key idea is to use dual information, such as spatial and spectral, or local and global features, in a complementary and effective way, and extract more accurate information from current iteration to use for the next iteration. In each DIR iteration, starting from current correspondence, a zoom-in process at each point is used to select well matched anchor pairs by a local mapping distortion criterion. These selected anchor pairs are then used to align spectral features (or other appropriate global features) whose dimension adaptively matches the capacity of the selected anchor pairs. Thanks to the effective combination of complementary information in a data-adaptive way, DIR is not only efficient but also robust to render accurate results within a few iterations. By choosing appropriate dual features, DIR has the flexibility to handle patch and partial matching as well. Extensive experiments on various data sets demonstrate the superiority of DIR over other state-of-the-art methods in terms of both accuracy and efficiency.

54.Mask2CAD: 3D Shape Prediction by Learning to Segment and Retrieve ⬇️

Object recognition has seen significant progress in the image domain, with focus primarily on 2D perception. We propose to leverage existing large-scale datasets of 3D models to understand the underlying 3D structure of objects seen in an image by constructing a CAD-based representation of the objects and their poses. We present Mask2CAD, which jointly detects objects in real-world images and for each detected object, optimizes for the most similar CAD model and its pose. We construct a joint embedding space between the detected regions of an image corresponding to an object and 3D CAD models, enabling retrieval of CAD models for an input RGB image. This produces a clean, lightweight representation of the objects in an image; this CAD-based representation ensures a valid, efficient shape representation for applications such as content creation or interactive scenarios, and makes a step towards understanding the transformation of real-world imagery to a synthetic domain. Experiments on real-world images from Pix3D demonstrate the advantage of our approach in comparison to state of the art. To facilitate future research, we additionally propose a new image-to-3D baseline on ScanNet which features larger shape diversity, real-world occlusions, and challenging image views.

55.Style is a Distribution of Features ⬇️

Neural style transfer (NST) is a powerful image generation technique that uses a convolutional neural network (CNN) to merge the content of one image with the style of another. Contemporary methods of NST use first or second order statistics of the CNN's features to achieve transfers with relatively little computational cost. However, these methods cannot fully extract the style from the CNN's features. We present a new algorithm for style transfer that fully extracts the style from the features by redefining the style loss as the Wasserstein distance between the distribution of features. Thus, we set a new standard in style transfer quality. In addition, we state two important interpretations of NST. The first is a re-emphasis from Li et al., which states that style is simply the distribution of features. The second states that NST is a type of generative adversarial network (GAN) problem.

56.HATNet: An End-to-End Holistic Attention Network for Diagnosis of Breast Biopsy Images ⬇️

Training end-to-end networks for classifying gigapixel size histopathological images is computationally intractable. Most approaches are patch-based and first learn local representations (patch-wise) before combining these local representations to produce image-level decisions. However, dividing large tissue structures into patches limits the context available to these networks, which may reduce their ability to learn representations from clinically relevant structures. In this paper, we introduce a novel attention-based network, the Holistic ATtention Network (HATNet) to classify breast biopsy images. We streamline the histopathological image classification pipeline and show how to learn representations from gigapixel size images end-to-end. HATNet extends the bag-of-words approach and uses self-attention to encode global information, allowing it to learn representations from clinically relevant tissue structures without any explicit supervision. It outperforms the previous best network Y-Net, which uses supervision in the form of tissue-level segmentation masks, by 8%. Importantly, our analysis reveals that HATNet learns representations from clinically relevant structures, and it matches the classification accuracy of human pathologists for this challenging test set. Our source code is available at \url{this https URL}

57.Robust and Generalizable Visual Representation Learning via Random Convolutions ⬇️

While successful for various computer vision tasks, deep neural networks have shown to be vulnerable to texture style shifts and small perturbations to which humans are robust. Hence, our goal is to train models in such a way that improves their robustness to these perturbations. We are motivated by the approximately shape-preserving property of randomized convolutions, which is due to distance preservation under random linear transforms. Intuitively, randomized convolutions create an infinite number of new domains with similar object shapes but random local texture. Therefore, we explore using outputs of multi-scale random convolutions as new images or mixing them with the original images during training. When applying a network trained with our approach to unseen domains, our method consistently improves the performance on domain generalization benchmarks and is scalable to ImageNet. Especially for the challenging scenario of generalizing to the sketch domain in PACS and to ImageNet-Sketch, our method outperforms state-of-art methods by a large margin. More interestingly, our method can benefit downstream tasks by providing a more robust pretrained visual representation.

58.GP-Aligner: Unsupervised Non-rigid Groupwise Point Set Registration Based On Optimized Group Latent Descriptor ⬇️

In this paper, we propose a novel method named GP-Aligner to deal with the problem of non-rigid groupwise point set registration. Compared to previous non-learning approaches, our proposed method gains competitive advantages by leveraging the power of deep neural networks to effectively and efficiently learn to align a large number of highly deformed 3D shapes with superior performance. Unlike most learning-based methods that use an explicit feature encoding network to extract the per-shape features and their correlations, our model leverages a model-free learnable latent descriptor to characterize the group relationship. More specifically, for a given group we first define an optimizable Group Latent Descriptor (GLD) to characterize the gruopwise relationship among a group of point sets. Each GLD is randomly initialized from a Gaussian distribution and then concatenated with the coordinates of each point of the associated point sets in the group. A neural network-based decoder is further constructed to predict the coherent drifts as the desired transformation from input groups of shapes to aligned groups of shapes. During the optimization process, GP-Aligner jointly updates all GLDs and weight parameters of the decoder network towards the minimization of an unsupervised groupwise alignment loss. After optimization, for each group our model coherently drives each point set towards a middle, common position (shape) without specifying one as the target. GP-Aligner does not require large-scale training data for network training and it can directly align groups of point sets in a one-stage optimization process. GP-Aligner shows both accuracy and computational efficiency improvement in comparison with state-of-the-art methods for groupwise point set registration. Moreover, GP-Aligner is shown great efficiency in aligning a large number of groups of real-world 3D shapes.

59.MRGAN: Multi-Rooted 3D Shape Generation with Unsupervised Part Disentanglement ⬇️

We present MRGAN, a multi-rooted adversarial network which generates part-disentangled 3D point-cloud shapes without part-based shape supervision. The network fuses multiple branches of tree-structured graph convolution layers which produce point clouds, with learnable constant inputs at the tree roots. Each branch learns to grow a different shape part, offering control over the shape generation at the part level. Our network encourages disentangled generation of semantic parts via two key ingredients: a root-mixing training strategy which helps decorrelate the different branches to facilitate disentanglement, and a set of loss terms designed with part disentanglement and shape semantics in mind. Of these, a novel convexity loss incentivizes the generation of parts that are more convex, as semantic parts tend to be. In addition, a root-dropping loss further ensures that each root seeds a single part, preventing the degeneration or over-growth of the point-producing branches. We evaluate the performance of our network on a number of 3D shape classes, and offer qualitative and quantitative comparisons to previous works and baseline approaches. We demonstrate the controllability offered by our part-disentangled generation through two applications for shape modeling: part mixing and individual part variation, without receiving segmented shapes as input.

60.Gradient Regularized Contrastive Learning for Continual Domain Adaptation ⬇️

Human beings can quickly adapt to environmental changes by leveraging learning experience. However, the poor ability of adapting to dynamic environments remains a major challenge for AI models. To better understand this issue, we study the problem of continual domain adaptation, where the model is presented with a labeled source domain and a sequence of unlabeled target domains. There are two major obstacles in this problem: domain shifts and catastrophic forgetting. In this work, we propose Gradient Regularized Contrastive Learning to solve the above obstacles. At the core of our method, gradient regularization plays two key roles: (1) enforces the gradient of contrastive loss not to increase the supervised training loss on the source domain, which maintains the discriminative power of learned features; (2) regularizes the gradient update on the new domain not to increase the classification loss on the old target domains, which enables the model to adapt to an in-coming target domain while preserving the performance of previously observed domains. Hence our method can jointly learn both semantically discriminative and domain-invariant features with labeled source domain and unlabeled target domains. The experiments on Digits, DomainNet and Office-Caltech benchmarks demonstrate the strong performance of our approach when compared to the state-of-the-art.

61.Video Super Resolution Based on Deep Learning: A comprehensive survey ⬇️

In recent years, deep learning has made great progress in the fields of image recognition, video analysis, natural language processing and speech recognition, including video super-resolution tasks. In this survey, we comprehensively investigate 28 state-of-the-art video super-resolution methods based on deep learning. It is well known that the leverage of information within video frames is important for video super-resolution. Hence we propose a taxonomy and classify the methods into six sub-categories according to the ways of utilizing inter-frame information. Moreover, the architectures and implementation details (including input and output, loss function and learning rate) of all the methods are depicted in details. Finally, we summarize and compare their performance on some benchmark datasets under different magnification factors. We also discuss some challenges, which need to be further addressed by researchers in the community of video super-resolution. Therefore, this work is expected to make a contribution to the future development of research in video super-resolution, and alleviate understandability and transferability of existing and future techniques into practice.

62.Crowdsourced 3D Mapping: A Combined Multi-View Geometry and Self-Supervised Learning Approach ⬇️

The ability to efficiently utilize crowdsourced visual data carries immense potential for the domains of large scale dynamic mapping and autonomous driving. However, state-of-the-art methods for crowdsourced 3D mapping assume prior knowledge of camera intrinsics. In this work, we propose a framework that estimates the 3D positions of semantically meaningful landmarks such as traffic signs without assuming known camera intrinsics, using only monocular color camera and GPS. We utilize multi-view geometry as well as deep learning based self-calibration, depth, and ego-motion estimation for traffic sign positioning, and show that combining their strengths is important for increasing the map coverage. To facilitate research on this task, we construct and make available a KITTI based 3D traffic sign ground truth positioning dataset. Using our proposed framework, we achieve an average single-journey relative and absolute positioning accuracy of 39cm and 1.26m respectively, on this dataset.

63.Approximated Bilinear Modules for Temporal Modeling ⬇️

We consider two less-emphasized temporal properties of video: 1. Temporal cues are fine-grained; 2. Temporal modeling needs reasoning. To tackle both problems at once, we exploit approximated bilinear modules (ABMs) for temporal modeling. There are two main points making the modules effective: two-layer MLPs can be seen as a constraint approximation of bilinear operations, thus can be used to construct deep ABMs in existing CNNs while reusing pretrained parameters; frame features can be divided into static and dynamic parts because of visual repetition in adjacent frames, which enables temporal modeling to be more efficient. Multiple ABM variants and implementations are investigated, from high performance to high efficiency. Specifically, we show how two-layer subnets in CNNs can be converted to temporal bilinear modules by adding an auxiliary-branch. Besides, we introduce snippet sampling and shifting inference to boost sparse-frame video classification performance. Extensive ablation studies are conducted to show the effectiveness of proposed techniques. Our models can outperform most state-of-the-art methods on Something-Something v1 and v2 datasets without Kinetics pretraining, and are also competitive on other YouTube-like action recognition datasets. Our code is available on this https URL.

64.Learning Disentangled Representations with Latent Variation Predictability ⬇️

Latent traversal is a popular approach to visualize the disentangled latent representations. Given a bunch of variations in a single unit of the latent representation, it is expected that there is a change in a single factor of variation of the data while others are fixed. However, this impressive experimental observation is rarely explicitly encoded in the objective function of learning disentangled representations. This paper defines the variation predictability of latent disentangled representations. Given image pairs generated by latent codes varying in a single dimension, this varied dimension could be closely correlated with these image pairs if the representation is well disentangled. Within an adversarial generation process, we encourage variation predictability by maximizing the mutual information between latent variations and corresponding image pairs. We further develop an evaluation metric that does not rely on the ground-truth generative factors to measure the disentanglement of latent representations. The proposed variation predictability is a general constraint that is applicable to the VAE and GAN frameworks for boosting disentanglement of latent representations. Experiments show that the proposed variation predictability correlates well with existing ground-truth-required metrics and the proposed algorithm is effective for disentanglement learning.

65.MirrorNet: Bio-Inspired Adversarial Attack for Camouflaged Object Segmentation ⬇️

Camouflaged objects are generally difficult to be detected in their natural environment even for human beings. In this paper, we propose a novel bio-inspired network, named the MirrorNet, that leverages both instance segmentation and adversarial attack for the camouflaged object segmentation. Differently from existing networks for segmentation, our proposed network possesses two segmentation streams: the main stream and the adversarial stream corresponding with the original image and its flipped image, respectively. The output from the adversarial stream is then fused into the main stream's result for the final camouflage map to boost up the segmentation accuracy. Extensive experiments conducted on the public CAMO dataset demonstrate the effectiveness of our proposed network. Our proposed method achieves 89% in accuracy, outperforming the state-of-the-arts.
Project Page: this https URL

66.Applying Semantic Segmentation to Autonomous Cars in the Snowy Environment ⬇️

This paper mainly focuses on environment perception in snowy situations which forms the backbone of the autonomous driving technology. For the purpose, semantic segmentation is employed to classify the objects while the vehicle is driven autonomously. We train the Fully Convolutional Networks (FCN) on our own dataset and present the experimental results. Finally, the outcomes are analyzed to give a conclusion. It can be concluded that the database still needs to be optimized and a favorable algorithm should be proposed to get better results.

67.OpenRooms: An End-to-End Open Framework for Photorealistic Indoor Scene Datasets ⬇️

Large-scale photorealistic datasets of indoor scenes, with ground truth geometry, materials and lighting, are important for deep learning applications in scene reconstruction and augmented reality. The associated shape, material and lighting assets can be scanned or artist-created, both of which are expensive; the resulting data is usually proprietary. We aim to make the dataset creation process for indoor scenes widely accessible, allowing researchers to transform casually acquired scans to large-scale datasets with high-quality ground truth. We achieve this by estimating consistent furniture and scene layout, ascribing high quality materials to all surfaces and rendering images with spatially-varying lighting consisting of area lights and environment maps. We demonstrate an instantiation of our approach on the publicly available ScanNet dataset. Deep networks trained on our proposed dataset achieve competitive performance for shape, material and lighting estimation on real images and can be used for photorealistic augmented reality applications, such as object insertion and material editing. Importantly, the dataset and all the tools to create such datasets from scans will be released, enabling others in the community to easily build large-scale datasets of their own. All code, models, data, dataset creation tool will be publicly released on our project page.

68.A Self-Training Approach for Point-Supervised Object Detection and Counting in Crowds ⬇️

In this paper, we propose a novel self-training approach which enables a typical object detector trained only with point-level annotations (i.e., objects are labeled with points) to estimate both the center points and sizes of crowded objects. Specifically, during training we utilize the available point annotations to directly supervise the estimation of the center points of objects. Based on a locally-uniform distribution assumption, we initialize pseudo object sizes from the point-level supervisory information, which are then leveraged to guide the regression of object sizes via a crowdedness-aware loss. Meanwhile, we propose a confidence and order-aware refinement scheme to continuously refine the initial pseudo object sizes such that the ability of the detector is increasingly boosted to simultaneously detect and count objects in crowds. Moreover, to address extremely crowded scenes, we propose an effective decoding method to improve the representation ability of the detector. Experimental results on the WiderFace benchmark show that our approach significantly outperforms state-of-the-art point-supervised methods under both detection and counting tasks, i.e., our method improves the average precision by more than 10% and reduces the counting error by 31.2%. In addition, our method obtains the best results on the dense crowd counting dataset (i.e., ShanghaiTech) and vehicle counting datasets (i.e., CARPK and PUCPR+) when compared with state-of-the-art counting-by-detection methods. We will make the code publicly available to facilitate future research.

69.Counting Fish and Dolphins in Sonar Images Using Deep Learning ⬇️

Deep learning provides the opportunity to improve upon conflicting reports considering the relationship between the Amazon river's fish and dolphin abundance and reduced canopy cover as a result of deforestation. Current methods of fish and dolphin abundance estimates are performed by on-site sampling using visual and capture/release strategies. We propose a novel approach to calculating fish abundance using deep learning for fish and dolphin estimates from sonar images taken from the back of a trolling boat. We consider a data set of 143 images ranging from 0-34 fish, and 0-3 dolphins provided by the Fund Amazonia research group. To overcome the data limitation, we test the capabilities of data augmentation on an unconventional 15/85 training/testing split. Using 20 training images, we simulate a gradient of data up to 25,000 images using augmented backgrounds and randomly placed/rotation cropped fish and dolphin taken from the training set. We then train four multitask network architectures: DenseNet201, InceptionNetV2, Xception, and MobileNetV2 to predict fish and dolphin numbers using two function approximation methods: regression and classification. For regression, Densenet201 performed best for fish and Xception best for dolphin with mean squared errors of 2.11 and 0.133 respectively. For classification, InceptionResNetV2 performed best for fish and MobileNetV2 best for dolphins with a mean error of 2.07 and 0.245 respectively. Considering the 123 testing images, our results show the success of data simulation for limited sonar data sets. We find DenseNet201 is able to identify dolphins after approximately 5000 training images, while fish required the full 25,000. Our method can be used to lower costs and expedite the data analysis of fish and dolphin abundance to real-time along the Amazon river and river systems worldwide.

70.Spatiotemporal Bundle Adjustment for Dynamic 3D Human Reconstruction in the Wild ⬇️

Bundle adjustment jointly optimizes camera intrinsics and extrinsics and 3D point triangulation to reconstruct a static scene. The triangulation constraint, however, is invalid for moving points captured in multiple unsynchronized videos and bundle adjustment is not designed to estimate the temporal alignment between cameras. We present a spatiotemporal bundle adjustment framework that jointly optimizes four coupled sub-problems: estimating camera intrinsics and extrinsics, triangulating static 3D points, as well as sub-frame temporal alignment between cameras and computing 3D trajectories of dynamic points. Key to our joint optimization is the careful integration of physics-based motion priors within the reconstruction pipeline, validated on a large motion capture corpus of human subjects. We devise an incremental reconstruction and alignment algorithm to strictly enforce the motion prior during the spatiotemporal bundle adjustment. This algorithm is further made more efficient by a divide and conquer scheme while still maintaining high accuracy. We apply this algorithm to reconstruct 3D motion trajectories of human bodies in dynamic events captured by multiple uncalibrated and unsynchronized video cameras in the wild. To make the reconstruction visually more interpretable, we fit a statistical 3D human body model to the asynchronous video streams.Compared to the baseline, the fitting significantly benefits from the proposed spatiotemporal bundle adjustment procedure. Because the videos are aligned with sub-frame precision, we reconstruct 3D motion at much higher temporal resolution than the input videos.

71.Dialog without Dialog Data: Learning Visual Dialog Agents from VQA Data ⬇️

Can we develop visually grounded dialog agents that can efficiently adapt to new tasks without forgetting how to talk to people? Such agents could leverage a larger variety of existing data to generalize to new tasks, minimizing expensive data collection and annotation. In this work, we study a setting we call "Dialog without Dialog", which requires agents to develop visually grounded dialog models that can adapt to new tasks without language level supervision. By factorizing intention and language, our model minimizes linguistic drift after fine-tuning for new tasks. We present qualitative results, automated metrics, and human studies that all show our model can adapt to new tasks and maintain language quality. Baselines either fail to perform well at new tasks or experience language drift, becoming unintelligible to humans. Code has been made available at this https URL

72.Hard negative examples are hard, but useful ⬇️

Triplet loss is an extremely common approach to distance metric learning. Representations of images from the same class are optimized to be mapped closer together in an embedding space than representations of images from different classes. Much work on triplet losses focuses on selecting the most useful triplets of images to consider, with strategies that select dissimilar examples from the same class or similar examples from different classes. The consensus of previous research is that optimizing with the \textit{hardest} negative examples leads to bad training behavior. That's a problem -- these hardest negatives are literally the cases where the distance metric fails to capture semantic similarity. In this paper, we characterize the space of triplets and derive why hard negatives make triplet loss training fail. We offer a simple fix to the loss function and show that, with this fix, optimizing with hard negative examples becomes feasible. This leads to more generalizable features, and image retrieval results that outperform state of the art for datasets with high intra-class variance.

73.Point Cloud Based Reinforcement Learning for Sim-to-Real and Partial Observability in Visual Navigation ⬇️

Reinforcement Learning (RL), among other learning-based methods, represents powerful tools to solve complex robotic tasks (e.g., actuation, manipulation, navigation, etc.), with the need for real-world data to train these systems as one of its most important limitations. The use of simulators is one way to address this issue, yet knowledge acquired in simulations does not work directly in the real-world, which is known as the sim-to-real transfer problem. While previous works focus on the nature of the images used as observations (e.g., textures and lighting), which has proven useful for a sim-to-sim transfer, they neglect other concerns regarding said observations, such as precise geometrical meanings, failing at robot-to-robot, and thus in sim-to-real transfers. We propose a method that learns on an observation space constructed by point clouds and environment randomization, generalizing among robots and simulators to achieve sim-to-real, while also addressing partial observability. We demonstrate the benefits of our methodology on the point goal navigation task, in which our method proves to be highly unaffected to unseen scenarios produced by robot-to-robot transfer, outperforms image-based baselines in robot-randomized experiments, and presents high performances in sim-to-sim conditions. Finally, we perform several experiments to validate the sim-to-real transfer to a physical domestic robot platform, confirming the out-of-the-box performance of our system.

74.MMDF: Mobile Microscopy Deep Framework ⬇️

In the last decade, a huge step was done in the field of mobile microscopes development as well as in the field of mobile microscopy application to real-life disease diagnostics and a lot of other important areas (air/water quality pollution, education, agriculture). In current study we applied image processing techniques from Deep Learning (in-focus/out-of-focus classification, image deblurring and denoising, multi-focus image fusion) to the data obtained from the mobile microscope. Overview of significant works for every task is presented, the most suitable approaches were highlighted. Chosen approaches were implemented as well as their performance were compared with classical computer vision techniques.

75.Cloud Detection through Wavelet Transforms in Machine Learning and Deep Learning ⬇️

Cloud detection is a specialized application of image recognition and object detection using remotely sensed data. The task presents a number of challenges, including analyzing images obtained in visible, infrared and multi-spectral frequencies, usually without ground truth data for comparison. Moreover, machine learning and deep learning (MLDL) algorithms applied to this task are required to be computationally efficient, as they are typically deployed in low-power devices and called to operate in real-time.
This paper explains Wavelet Transform (WT) theory, comparing it to more widely used image and signal processing transforms, and explores the use of WT as a powerful signal compressor and feature extractor for MLDL classifiers.

76.Towards Learning Convolutions from Scratch ⬇️

Convolution is one of the most essential components of architectures used in computer vision. As machine learning moves towards reducing the expert bias and learning it from data, a natural next step seems to be learning convolution-like structures from scratch. This, however, has proven elusive. For example, current state-of-the-art architecture search algorithms use convolution as one of the existing modules rather than learning it from data. In an attempt to understand the inductive bias that gives rise to convolutions, we investigate minimum description length as a guiding principle and show that in some settings, it can indeed be indicative of the performance of architectures. To find architectures with small description length, we propose $\beta$-LASSO, a simple variant of LASSO algorithm that, when applied on fully-connected networks for image classification tasks, learns architectures with local connections and achieves state-of-the-art accuracies for training fully-connected nets on CIFAR-10 (85.19%), CIFAR-100 (59.56%) and SVHN (94.07%) bridging the gap between fully-connected and convolutional nets.

77.Orpheus: A New Deep Learning Framework for Easy Deployment and Evaluation of Edge Inference ⬇️

Optimising deep learning inference across edge devices and optimisation targets such as inference time, memory footprint and power consumption is a key challenge due to the ubiquity of neural networks. Today, production deep learning frameworks provide useful abstractions to aid machine learning engineers and systems researchers. However, in exchange they can suffer from compatibility challenges (especially on constrained platforms), inaccessible code complexity, or design choices that otherwise limit research from a systems perspective. This paper presents Orpheus, a new deep learning framework for easy prototyping, deployment and evaluation of inference optimisations. Orpheus features a small codebase, minimal dependencies, and a simple process for integrating other third party systems. We present some preliminary evaluation results.

78.Hardware Implementation of Hyperbolic Tangent Function using Catmull-Rom Spline Interpolation ⬇️

Deep neural networks yield the state of the art results in many computer vision and human machine interface tasks such as object recognition, speech recognition etc. Since, these networks are computationally expensive, customized accelerators are designed for achieving the required performance at lower cost and power. One of the key building blocks of these neural networks is non-linear activation function such as sigmoid, hyperbolic tangent (tanh), and ReLU. A low complexity accurate hardware implementation of the activation function is required to meet the performance and area targets of the neural network accelerators. This paper presents an implementation of tanh function using the Catmull-Rom spline interpolation. State of the art results are achieved using this method with comparatively smaller logic area.

79.Attention-based Graph ResNet for Motor Intent Detection from Raw EEG signals ⬇️

In previous studies, decoding electroencephalography (EEG) signals has not considered the topological relationship of EEG electrodes. However, the latest neuroscience has suggested brain network connectivity. Thus, the exhibited interaction between EEG channels might not be appropriately measured via Euclidean distance. To fill the gap, an attention-based graph residual network, a novel structure of Graph Convolutional Neural Network (GCN), was presented to detect human motor intents from raw EEG signals, where the topological structure of EEG electrodes was built as a graph. Meanwhile, deep residual learning with a full-attention architecture was introduced to address the degradation problem concerning deeper networks in raw EEG motor imagery (MI) data. Individual variability, the critical and longstanding challenge underlying EEG signals, has been successfully handled with the state-of-the-art performance, 98.08% accuracy at the subject level, 94.28% for 20 subjects. Numerical results were promising that the implementation of the graph-structured topology was superior to decode raw EEG data. The innovative deep learning approach was expected to entail a universal method towards both neuroscience research and real-world EEG-based practical applications, e.g., seizure prediction.

80.Image-driven discriminative and generative machine learning algorithms for establishing microstructure-processing relationships ⬇️

We investigate methods of microstructure representation for the purpose of predicting processing condition from microstructure image data. A binary alloy (uranium-molybdenum) that is currently under development as a nuclear fuel was studied for the purpose of developing an improved machine learning approach to image recognition, characterization, and building predictive capabilities linking microstructure to processing conditions. Here, we test different microstructure representations and evaluate model performance based on the F1 score. A F1 score of 95.1% was achieved for distinguishing between micrographs corresponding to ten different thermo-mechanical material processing conditions. We find that our newly developed microstructure representation describes image data well, and the traditional approach of utilizing area fractions of different phases is insufficient for distinguishing between multiple classes using a relatively small, imbalanced original data set of 272 images. To explore the applicability of generative methods for supplementing such limited data sets, generative adversarial networks were trained to generate artificial microstructure images. Two different generative networks were trained and tested to assess performance. Challenges and best practices associated with applying machine learning to limited microstructure image data sets is also discussed. Our work has implications for quantitative microstructure analysis, and development of microstructure-processing relationships in limited data sets typical of metallurgical process design studies.

81.XCAT-GAN for Synthesizing 3D Consistent Labeled Cardiac MR Images on Anatomically Variable XCAT Phantoms ⬇️

Generative adversarial networks (GANs) have provided promising data enrichment solutions by synthesizing high-fidelity images. However, generating large sets of labeled images with new anatomical variations remains unexplored. We propose a novel method for synthesizing cardiac magnetic resonance (CMR) images on a population of virtual subjects with a large anatomical variation, introduced using the 4D eXtended Cardiac and Torso (XCAT) computerized human phantom. We investigate two conditional image synthesis approaches grounded on a semantically-consistent mask-guided image generation technique: 4-class and 8-class XCAT-GANs. The 4-class technique relies on only the annotations of the heart; while the 8-class technique employs a predicted multi-tissue label map of the heart-surrounding organs and provides better guidance for our conditional image synthesis. For both techniques, we train our conditional XCAT-GAN with real images paired with corresponding labels and subsequently at the inference time, we substitute the labels with the XCAT derived ones. Therefore, the trained network accurately transfers the tissue-specific textures to the new label maps. By creating 33 virtual subjects of synthetic CMR images at the end-diastolic and end-systolic phases, we evaluate the usefulness of such data in the downstream cardiac cavity segmentation task under different augmentation strategies. Results demonstrate that even with only 20% of real images (40 volumes) seen during training, segmentation performance is retained with the addition of synthetic CMR images. Moreover, the improvement in utilizing synthetic images for augmenting the real data is evident through the reduction of Hausdorff distance up to 28% and an increase in the Dice score up to 5%, indicating a higher similarity to the ground truth in all dimensions.

82.ALF: Autoencoder-based Low-rank Filter-sharing for Efficient Convolutional Neural Networks ⬇️

Closing the gap between the hardware requirements of state-of-the-art convolutional neural networks and the limited resources constraining embedded applications is the next big challenge in deep learning research. The computational complexity and memory footprint of such neural networks are typically daunting for deployment in resource constrained environments. Model compression techniques, such as pruning, are emphasized among other optimization methods for solving this problem. Most existing techniques require domain expertise or result in irregular sparse representations, which increase the burden of deploying deep learning applications on embedded hardware accelerators. In this paper, we propose the autoencoder-based low-rank filter-sharing technique technique (ALF). When applied to various networks, ALF is compared to state-of-the-art pruning methods, demonstrating its efficient compression capabilities on theoretical metrics as well as on an accurate, deterministic hardware-model. In our experiments, ALF showed a reduction of 70% in network parameters, 61% in operations and 41% in execution time, with minimal loss in accuracy.

83.Dual Distribution Alignment Network for Generalizable Person Re-Identification ⬇️

Domain generalization (DG) serves as a promising solution to handle person Re-Identification (Re-ID), which trains the model using labels from the source domain alone, and then directly adopts the trained model to the target domain without model updating. However, existing DG approaches are usually disturbed by serious domain variations due to significant dataset variations. Subsequently, DG highly relies on designing domain-invariant features, which is however not well exploited, since most existing approaches directly mix multiple datasets to train DG based models without considering the local dataset similarities, i.e., examples that are very similar but from different domains. In this paper, we present a Dual Distribution Alignment Network (DDAN), which handles this challenge by mapping images into a domain-invariant feature space by selectively aligning distributions of multiple source domains. Such an alignment is conducted by dual-level constraints, i.e., the domain-wise adversarial feature learning and the identity-wise similarity enhancement. We evaluate our DDAN on a large-scale Domain Generalization Re-ID (DG Re-ID) benchmark. Quantitative results demonstrate that the proposed DDAN can well align the distributions of various source domains, and significantly outperforms all existing domain generalization approaches.

84.Uniformizing Techniques to Process CT scans with 3D CNNs for Tuberculosis Prediction ⬇️

A common approach to medical image analysis on volumetric data uses deep 2D convolutional neural networks (CNNs). This is largely attributed to the challenges imposed by the nature of the 3D data: variable volume size, GPU exhaustion during optimization. However, dealing with the individual slices independently in 2D CNNs deliberately discards the depth information which results in poor performance for the intended task. Therefore, it is important to develop methods that not only overcome the heavy memory and computation requirements but also leverage the 3D information. To this end, we evaluate a set of volume uniformizing methods to address the aforementioned issues. The first method involves sampling information evenly from a subset of the volume. Another method exploits the full geometry of the 3D volume by interpolating over the z-axis. We demonstrate performance improvements using controlled ablation studies as well as put this approach to the test on the ImageCLEF Tuberculosis Severity Assessment 2019 benchmark. We report 73% area under curve (AUC) and binary classification accuracy (ACC) of 67.5% on the test set beating all methods which leveraged only image information (without using clinical meta-data) achieving 5-th position overall. All codes and models are made available at this https URL.

85.UIAI System for Short-Duration Speaker Verification Challenge 2020 ⬇️

In this work, we present the system description of the UIAI entry for the short-duration speaker verification (SdSV) challenge 2020. Our focus is on Task 1 dedicated to text-dependent speaker verification. We investigate different feature extraction and modeling approaches for automatic speaker verification (ASV) and utterance verification (UV). We have also studied different fusion strategies for combining UV and ASV modules. Our primary submission to the challenge is the fusion of seven subsystems which yields a normalized minimum detection cost function (minDCF) of 0.072 and an equal error rate (EER) of 2.14% on the evaluation set. The single system consisting of a pass-phrase identification based model with phone-discriminative bottleneck features gives a normalized minDCF of 0.118 and achieves 19% relative improvement over the state-of-the-art challenge baseline.

86.Regularized Flexible Activation Function Combinations for Deep Neural Networks ⬇️

Activation in deep neural networks is fundamental to achieving non-linear mappings. Traditional studies mainly focus on finding fixed activations for a particular set of learning tasks or model architectures. The research on flexible activation is quite limited in both designing philosophy and application scenarios. In this study, three principles of choosing flexible activation components are proposed and a general combined form of flexible activation functions is implemented. Based on this, a novel family of flexible activation functions that can replace sigmoid or tanh in LSTM cells are implemented, as well as a new family by combining ReLU and ELUs. Also, two new regularisation terms based on assumptions as prior knowledge are introduced. It has been shown that LSTM models with proposed flexible activations P-Sig-Ramp provide significant improvements in time series forecasting, while the proposed P-E2-ReLU achieves better and more stable performance on lossy image compression tasks with convolutional auto-encoders. In addition, the proposed regularization terms improve the convergence, performance and stability of the models with flexible activation functions.

87.MACU-Net Semantic Segmentation from High-Resolution Remote Sensing Images ⬇️

Semantic segmentation of remote sensing images plays an important role in land resource management, yield estimation, and economic assessment. U-Net is a sophisticated encoder-decoder architecture which has been frequently used in medical image segmentation and has attained prominent performance. And asymmetric convolution block can enhance the square convolution kernels using asymmetric convolutions. In this paper, based on U-Net and asymmetric convolution block, we incorporate multi-scale features generated by different layers of U-Net and design a multi-scale skip connected architecture, MACU-Net, for semantic segmentation using high-resolution remote sensing images. Our design has the following advantages: (1) The multi-scale skip connections combine and realign semantic features contained both in low-level and high-level feature maps with different scales; (2) the asymmetric convolution block strengthens the representational capacity of a standard convolution layer. Experiments conducted on two remote sensing image datasets captured by separate satellites demonstrate that the performance of our MACU-Net transcends the U-Net, SegNet, DeepLab V3+, and other baseline algorithms.

88.A Preliminary Exploration into an Alternative CellLineNet: An Evolutionary Approach ⬇️

Within this paper, the exploration of an evolutionary approach to an alternative CellLineNet: a convolutional neural network adept at the classification of epithelial breast cancer cell lines, is presented. This evolutionary algorithm introduces control variables that guide the search of architectures in the search space of inverted residual blocks, bottleneck blocks, residual blocks and a basic 2x2 convolutional block. The promise of EvoCELL is predicting what combination or arrangement of the feature extracting blocks that produce the best model architecture for a given task. Therein, the performance of how the fittest model evolved after each generation is shown. The final evolved model CellLineNet V2 classifies 5 types of epithelial breast cell lines consisting of two human cancer lines, 2 normal immortalized lines, and 1 immortalized mouse line (MDA-MB-468, MCF7, 10A, 12A and HC11). The Multiclass Cell Line Classification Convolutional Neural Network extends our earlier work on a Binary Breast Cancer Cell Line Classification model. This paper presents an on-going exploratory approach to neural network architecture design and is presented for further study.

89.Tighter risk certificates for neural networks ⬇️

This paper presents empirical studies regarding training probabilistic neural networks using training objectives derived from PAC-Bayes bounds. In the context of probabilistic neural networks, the output of training is a probability distribution over network weights. We present two training objectives, used here for the first time in connection with training neural networks. These two training objectives are derived from tight PAC-Bayes bounds, one of which is new. We also re-implement a previously used training objective based on a classical PAC-Bayes bound, to compare the properties of the predictors learned using the different training objectives. We compute risk certificates that are valid on any unseen examples for the learnt predictors. We further experiment with different types of priors on the weights (both data-free and data-dependent priors) and neural network architectures. Our experiments on MNIST and CIFAR-10 show that our training methods produce competitive test set errors and non-vacuous risk bounds with much tighter values than previous results in the literature, showing promise not only to guide the learning algorithm through bounding the risk but also for model selection. These observations suggest that the methods studied here might be good candidates for self-bounding learning.

90.CNN Detection of GAN-Generated Face Images based on Cross-Band Co-occurrences Analysis ⬇️

Last-generation GAN models allow to generate synthetic images which are visually indistinguishable from natural ones, raising the need to develop tools to distinguish fake and natural images thus contributing to preserve the trustworthiness of digital images. While modern GAN models can generate very high-quality images with no visible spatial artifacts, reconstruction of consistent relationships among colour channels is expectedly more difficult. In this paper, we propose a method for distinguishing GAN-generated from natural images by exploiting inconsistencies among spectral bands, with specific focus on the generation of synthetic face images. Specifically, we use cross-band co-occurrence matrices, in addition to spatial co-occurrence matrices, as input to a CNN model, which is trained to distinguish between real and synthetic faces. The results of our experiments confirm the goodness of our approach which outperforms a similar detection technique based on intra-band spatial co-occurrences only. The performance gain is particularly significant with regard to robustness against post-processing, like geometric transformations, filtering and contrast manipulations.

91.3D Neural Network for Lung Cancer Risk Prediction on CT Volumes ⬇️

With an estimated 160,000 deaths in 2018, lung cancer is the most common cause of cancer death in the United States. Lung cancer CT screening has been shown to reduce mortality by up to 40% and is now included in US screening guidelines. Reducing the high error rates in lung cancer screening is imperative because of the high clinical and financial costs caused by diagnosis mistakes. Despite the use of standards for radiological diagnosis, persistent inter-grader variability and incomplete characterization of comprehensive imaging findings remain as limitations of current methods. These limitations suggest opportunities for more sophisticated systems to improve performance and inter-reader consistency. In this report, we reproduce a state-of-the-art deep learning algorithm for lung cancer risk prediction. Our model predicts malignancy probability and risk bucket classification from lung CT studies. This allows for risk categorization of patients being screened and suggests the most appropriate surveillance and management. Combining our solution high accuracy, consistency and fully automated nature, our approach may enable highly efficient screening procedures and accelerate the adoption of lung cancer screening.

92.Modal Uncertainty Estimation via Discrete Latent Representation ⬇️

Many important problems in the real world don't have unique solutions. It is thus important for machine learning models to be capable of proposing different plausible solutions with meaningful probability measures. In this work we introduce such a deep learning framework that learns the one-to-many mappings between the inputs and outputs, together with faithful uncertainty measures. We call our framework {\it modal uncertainty estimation} since we model the one-to-many mappings to be generated through a set of discrete latent variables, each representing a latent mode hypothesis that explains the corresponding type of input-output relationship. The discrete nature of the latent representations thus allows us to estimate for any input the conditional probability distribution of the outputs very effectively. Both the discrete latent space and its uncertainty estimation are jointly learned during training. We motivate our use of discrete latent space through the multi-modal posterior collapse problem in current conditional generative models, then develop the theoretical background, and extensively validate our method on both synthetic and realistic tasks. Our framework demonstrates significantly more accurate uncertainty estimation than the current state-of-the-art methods, and is informative and convenient for practical use.

93.Joint Featurewise Weighting and Lobal Structure Learning for Multi-view Subspace Clustering ⬇️

Multi-view clustering integrates multiple feature sets, which reveal distinct aspects of the data and provide complementary information to each other, to improve the clustering performance. It remains challenging to effectively exploit complementary information across multiple views since the original data often contain noise and are highly redundant. Moreover, most existing multi-view clustering methods only aim to explore the consistency of all views while ignoring the local structure of each view. However, it is necessary to take the local structure of each view into consideration, because different views would present different geometric structures while admitting the same cluster structure. To address the above issues, we propose a novel multi-view subspace clustering method via simultaneously assigning weights for different features and capturing local information of data in view-specific self-representation feature spaces. Especially, a common cluster structure regularization is adopted to guarantee consistency among different views. An efficient algorithm based on an augmented Lagrangian multiplier is also developed to solve the associated optimization problem. Experiments conducted on several benchmark datasets demonstrate that the proposed method achieves state-of-the-art performance. We provide the Matlab code on this https URL.

94.All-Optical Information Processing Capacity of Diffractive Surfaces ⬇️

Precise engineering of materials and surfaces has been at the heart of some of the recent advances in optics and photonics. These advances around the engineering of materials with new functionalities have also opened up exciting avenues for designing trainable surfaces that can perform computation and machine learning tasks through light-matter interaction and diffraction. Here, we analyze the information processing capacity of coherent optical networks formed by diffractive surfaces that are trained to perform an all-optical computational task between a given input and output field-of-view. We prove that the dimensionality of the all-optical solution space covering the complex-valued transformations between the input and output fields-of-view is linearly proportional to the number of diffractive surfaces within the optical network, up to a limit that is dictated by the extent of the input and output fields-of-view. Deeper diffractive networks that are composed of larger numbers of trainable surfaces can cover a higher dimensional subspace of the complex-valued linear transformations between a larger input field-of-view and a larger output field-of-view, and exhibit depth advantages in terms of their statistical inference, learning and generalization capabilities for different image classification tasks, when compared with a single trainable diffractive surface. These analyses and conclusions are broadly applicable to various forms of diffractive surfaces, including e.g., plasmonic and/or dielectric-based metasurfaces and flat optics that can be used to form all-optical processors.

95.Selection of Proper EEG Channels for Subject Intention Classification Using Deep Learning ⬇️

Brain signals could be used to control devices to assist individuals with disabilities. Signals such as electroencephalograms are complicated and hard to interpret. A set of signals are collected and should be classified to identify the intention of the subject. Different approaches have tried to reduce the number of channels before sending them to a classifier. We are proposing a deep learning-based method for selecting an informative subset of channels that produce high classification accuracy. The proposed network could be trained for an individual subject for the selection of an appropriate set of channels. Reduction of the number of channels could reduce the complexity of brain-computer-interface devices. Our method could find a subset of channels. The accuracy of our approach is comparable with a model trained on all channels. Hence, our model's temporal and power costs are low, while its accuracy is kept high.