20200225.md

ArXiv cs.CV --Tue, 25 Feb 2020

1.Maximum Entropy on the Mean: A Paradigm Shift for Regularization in Image Deblurring ⬇️

Image deblurring is a notoriously challenging ill-posed inverse problem. In recent years, a wide variety of approaches have been proposed based upon regularization at the level of the image or on techniques from machine learning. We propose an alternative approach, shifting the paradigm towards regularization at the level of the probability distribution on the space of images. Our method is based upon the idea of maximum entropy on the mean wherein we work at the level of the probability density function of the image whose expectation is our estimate of the ground truth. Using techniques from convex analysis and probability theory, we show that the method is computationally feasible and amenable to very large blurs. Moreover, when images are imbedded with symbology (a known pattern), we show how our method can be applied to approximate the unknown blur kernel with remarkable effects. While our method is stable with respect to small amounts of noise, it does not actively denoise. However, for moderate to large amounts of noise, it performs well by preconditioned denoising with a state of the art method.

2.Boosting rare benthic macroinvertebrates taxa identification with one-class classification ⬇️

Insect monitoring is crucial for understanding the consequences of rapid ecological changes, but taxa identification currently requires tedious manual expert work and cannot be scaled-up efficiently. Deep convolutional neural networks (CNNs), provide a viable way to significantly increase the biomonitoring volumes. However, taxa abundances are typically very imbalanced and the amounts of training images for the rarest classes are simply too low for deep CNNs. As a result, the samples from the rare classes are often completely missed, while detecting them has biological importance. In this paper, we propose combining the trained deep CNN with one-class classifiers to improve the rare species identification. One-class classification models are traditionally trained with much fewer samples and they can provide a mechanism to indicate samples potentially belonging to the rare classes for human inspection. Our experiments confirm that the proposed approach may indeed support moving towards partial automation of the taxa identification task.

3.Suppressing Uncertainties for Large-Scale Facial Expression Recognition ⬇️

Annotating a qualitative large-scale facial expression dataset is extremely difficult due to the uncertainties caused by ambiguous facial expressions, low-quality facial images, and the subjectiveness of annotators. These uncertainties lead to a key challenge of large-scale Facial Expression Recognition (FER) in deep learning era. To address this problem, this paper proposes a simple yet efficient Self-Cure Network (SCN) which suppresses the uncertainties efficiently and prevents deep networks from over-fitting uncertain facial images. Specifically, SCN suppresses the uncertainty from two different aspects: 1) a self-attention mechanism over mini-batch to weight each training sample with a ranking regularization, and 2) a careful relabeling mechanism to modify the labels of these samples in the lowest-ranked group. Experiments on synthetic FER datasets and our collected WebEmotion dataset validate the effectiveness of our method. Results on public benchmarks demonstrate that our SCN outperforms current state-of-the-art methods with \textbf{88.14}% on RAF-DB, \textbf{60.23}% on AffectNet, and \textbf{89.35}% on FERPlus. The code will be available at \href{this https URL}{this https URL}.

4.Sketchformer: Transformer-based Representation for Sketched Structure ⬇️

Sketchformer is a novel transformer-based representation for encoding free-hand sketches input in a vector form, i.e. as a sequence of strokes. Sketchformer effectively addresses multiple tasks: sketch classification, sketch based image retrieval (SBIR), and the reconstruction and interpolation of sketches. We report several variants exploring continuous and tokenized input representations, and contrast their performance. Our learned embedding, driven by a dictionary learning tokenization scheme, yields state of the art performance in classification and image retrieval tasks, when compared against baseline representations driven by LSTM sequence to sequence architectures: SketchRNN and derivatives. We show that sketch reconstruction and interpolation are improved significantly by the Sketchformer embedding for complex sketches with longer stroke sequences.

5.Joint Learning of Assignment and Representation for Biometric Group Membership ⬇️

This paper proposes a framework for group membership protocols preventing the curious but honest server from reconstructing the enrolled biometric signatures and inferring the identity of querying clients. This framework learns the embedding parameters, group representations and assignments simultaneously. Experiments show the trade-off between security/privacy and verification/identification performances.

6.Comparing View-Based and Map-Based Semantic Labelling in Real-Time SLAM ⬇️

Generally capable Spatial AI systems must build persistent scene representations where geometric models are combined with meaningful semantic labels. The many approaches to labelling scenes can be divided into two clear groups: view-based which estimate labels from the input view-wise data and then incrementally fuse them into the scene model as it is built; and map-based which label the generated scene model. However, there has so far been no attempt to quantitatively compare view-based and map-based labelling. Here, we present an experimental framework and comparison which uses real-time height map fusion as an accessible platform for a fair comparison, opening up the route to further systematic research in this area.

7.Guessing State Tracking for Visual Dialogue ⬇️

The Guesser plays an important role in GuessWhat?! like visual dialogues. It locates the target object in an image supposed by an oracle oneself over a question-answer based dialogue between a Questioner and the Oracle. Most existing guessers make one and only one guess after receiving all question-answer pairs in a dialogue with predefined number of rounds. This paper proposes the guessing state for the guesser, and regards guess as a process with change of guessing state through a dialogue. A guessing state tracking based guess model is therefore proposed. The guessing state is defined as a distribution on candidate objects in the image. A state update algorithm including three modules is given. UoVR updates the representation of the image according to current guessing state, QAEncoder encodes the question-answer pairs, and UoGS updates the guessing state by combining both information from the image and dialogue history. With the guessing state in hand, two loss functions are defined as supervisions for model training. Early supervision brings supervision to guesser at early rounds, and incremental supervision brings monotonicity to the guessing state. Experimental results on GuessWhat?! dataset show that our model significantly outperforms previous models, achieves new state-of-the-art, especially, the success rate of guessing 83.3% is approaching human-level performance 84.4%.

8.Anatomy-aware 3D Human Pose Estimation in Videos ⬇️

In this work, we propose a new solution for 3D human pose estimation in videos. Instead of directly regressing the 3D joint locations, we draw inspiration from the human skeleton anatomy and decompose the task into bone direction prediction and bone length prediction, from which the 3D joint locations can be completely derived. Our motivation is the fact that the bone lengths of a human skeleton remain consistent across time. This promotes us to develop effective techniques to utilize global information across {\it all} the frames in a video for high-accuracy bone length prediction. Moreover, for the bone direction prediction network, we propose a fully-convolutional propagating architecture with long skip connections. Essentially, it predicts the directions of different bones hierarchically without using any time-consuming memory units (e.g. LSTM). A novel joint shift loss is further introduced to bridge the training of the bone length and bone direction prediction networks. Finally, we employ an implicit attention mechanism to feed the 2D keypoint visibility scores into the model as extra guidance, which significantly mitigates the depth ambiguity in many challenging poses. Our full model outperforms the previous best results on Human3.6M and MPI-INF-3DHP datasets, where comprehensive evaluation validates the effectiveness of our model.

9.Sketch Less for More: On-the-Fly Fine-Grained Sketch Based Image Retrieval ⬇️

Fine-grained sketch-based image retrieval (FG-SBIR) addresses the problem of retrieving a particular photo instance given a user's query sketch. Its widespread applicability is however hindered by the fact that drawing a sketch takes time, and most people struggle to draw a complete and faithful sketch. In this paper, we reformulate the conventional FG-SBIR framework to tackle these challenges, with the ultimate goal of retrieving the target photo with the least number of strokes possible. We further propose an on-the-fly design that starts retrieving as soon as the user starts drawing. To accomplish this, we devise a reinforcement learning-based cross-modal retrieval framework that directly optimizes rank of the ground-truth photo over a complete sketch drawing episode. Additionally, we introduce a novel reward scheme that circumvents the problems related to irrelevant sketch strokes, and thus provides us with a more consistent rank list during the retrieval. We achieve superior early-retrieval efficiency over state-of-the-art methods and alternative baselines on two publicly available fine-grained sketch retrieval datasets.

10.Uncertainty based Class Activation Maps for Visual Question Answering ⬇️

Understanding and explaining deep learning models is an imperative task. Towards this, we propose a method that obtains gradient-based certainty estimates that also provide visual attention maps. Particularly, we solve for visual question answering task. We incorporate modern probabilistic deep learning methods that we further improve by using the gradients for these estimates. These have two-fold benefits: a) improvement in obtaining the certainty estimates that correlate better with misclassified samples and b) improved attention maps that provide state-of-the-art results in terms of correlation with human attention regions. The improved attention maps result in consistent improvement for various methods for visual question answering. Therefore, the proposed technique can be thought of as a recipe for obtaining improved certainty estimates and explanations for deep learning models. We provide detailed empirical analysis for the visual question answering task on all standard benchmarks and comparison with state of the art methods.

11.PUGeo-Net: A Geometry-centric Network for 3D Point Cloud Upsampling ⬇️

This paper addresses the problem of generating uniform dense point clouds to describe the underlying geometric structures from given sparse point clouds. Due to the irregular and unordered nature, point cloud densification as a generative task is challenging. To tackle the challenge, we propose a novel deep neural network based method, called PUGeo-Net, that learns a $3\times 3$ linear transformation matrix $\bf T$ for each input point. Matrix $\mathbf T$ approximates the augmented Jacobian matrix of a local parameterization and builds a one-to-one correspondence between the 2D parametric domain and the 3D tangent plane so that we can lift the adaptively distributed 2D samples (which are also learned from data) to 3D space. After that, we project the samples to the curved surface by computing a displacement along the normal of the tangent plane. PUGeo-Net is fundamentally different from the existing deep learning methods that are largely motivated by the image super-resolution techniques and generate new points in the abstract feature space. Thanks to its geometry-centric nature, PUGeo-Net works well for both CAD models with sharp features and scanned models with rich geometric details. Moreover, PUGeo-Net can compute the normal for the original and generated points, which is highly desired by the surface reconstruction algorithms. Computational results show that PUGeo-Net, the first neural network that can jointly generate vertex coordinates and normals, consistently outperforms the state-of-the-art in terms of accuracy and efficiency for upsampling factor $4\sim 16$

12.Automatic Estimation of Sphere Centers from Images of Calibrated Cameras ⬇️

Calibration of devices with different modalities is a key problem in robotic vision. Regular spatial objects, such as planes, are frequently used for this task. This paper deals with the automatic detection of ellipses in camera images, as well as to estimate the 3D position of the spheres corresponding to the detected 2D ellipses. We propose two novel methods to (i) detect an ellipse in camera images and (ii) estimate the spatial location of the corresponding sphere if its size is known. The algorithms are tested both quantitatively and qualitatively. They are applied for calibrating the sensor system of autonomous cars equipped with digital cameras, depth sensors and LiDAR devices.

13.On the General Value of Evidence, and Bilingual Scene-Text Visual Question Answering ⬇️

Visual Question Answering (VQA) methods have made incredible progress, but suffer from a failure to generalize. This is visible in the fact that they are vulnerable to learning coincidental correlations in the data rather than deeper relations between image content and ideas expressed in language. We present a dataset that takes a step towards addressing this problem in that it contains questions expressed in two languages, and an evaluation process that co-opts a well understood image-based metric to reflect the method's ability to reason. Measuring reasoning directly encourages generalization by penalizing answers that are coincidentally correct. The dataset reflects the scene-text version of the VQA problem, and the reasoning evaluation can be seen as a text-based version of a referring expression challenge. Experiments and analysis are provided that show the value of the dataset.

14.Mnemonics Training: Multi-Class Incremental Learning without Forgetting ⬇️

Multi-Class Incremental Learning (MCIL) aims to learn new concepts by incrementally updating a model trained on previous concepts. However, there is an inherent trade-off to effectively learning new concepts without forgetting previous ones, potentially leading to catastrophic forgetting of previous concepts. To alleviate this issue, it has been proposed to keep around a few examples of the previous concepts but the effectiveness of this approach heavily depends on the representativeness of these examples. This paper proposes a novel and automatic framework we call mnemonics, where we parameterize exemplars and make them optimizable in an end-to-end manner. We train the framework through bilevel optimizations, i.e., model-level and exemplar-level. We conduct extensive experiments on three MCIL benchmarks, CIFAR-100, ImageNet-Subset and ImageNet, and show that using mnemonics exemplars can surpass the state-of-the-art by a large margin. Interestingly and quite intriguingly, the mnemonics exemplars tend to be on the boundaries between classes.

15.ABCNet: Real-time Scene Text Spotting with Adaptive Bezier-Curve Network ⬇️

Scene text detection and recognition has received increasing research attention. Existing methods can be roughly categorized into two groups: character-based and segmentation-based. These methods either are costly for character annotation or need to maintain a complex pipeline, which is often not suitable for real-time applications. Here we address the problem by proposing the Adaptive Bezier-Curve Network (ABCNet). Our contributions are three-fold: 1) For the first time, we adaptively fit arbitrarily-shaped text by a parameterized Bezier curve. 2) We design a novel BezierAlign layer for extracting accurate convolution features of a text instance with arbitrary shapes, significantly improving the precision compared with previous methods. 3) Compared with standard bounding box detection, our Bezier curve detection introduces negligible computation overhead, resulting in superiority of our method in both efficiency and accuracy. Experiments on arbitrarily-shaped benchmark datasets, namely Total-Text and CTW1500, demonstrate that ABCNet achieves state-of-the-art accuracy, meanwhile significantly improving the speed. In particular, on Total-Text, our realtime version is over 10 times faster than recent state-of-the-art methods with a competitive recognition accuracy. Code is available at this https URL

16.Learning Attentive Pairwise Interaction for Fine-Grained Classification ⬇️

Fine-grained classification is a challenging problem, due to subtle differences among highly-confused categories. Most approaches address this difficulty by learning discriminative representation of individual input image. On the other hand, humans can effectively identify contrastive clues by comparing image pairs. Inspired by this fact, this paper proposes a simple but effective Attentive Pairwise Interaction Network (API-Net), which can progressively recognize a pair of fine-grained images by interaction. Specifically, API-Net first learns a mutual feature vector to capture semantic differences in the input pair. It then compares this mutual vector with individual vectors to generate gates for each input image. These distinct gate vectors inherit mutual context on semantic differences, which allow API-Net to attentively capture contrastive clues by pairwise interaction between two images. Additionally, we train API-Net in an end-to-end manner with a score ranking regularization, which can further generalize API-Net by taking feature priorities into account. We conduct extensive experiments on five popular benchmarks in fine-grained classification. API-Net outperforms the recent SOTA methods, i.e., CUB-200-2011 (90.0%), Aircraft(93.9%), Stanford Cars (95.3%), Stanford Dogs (90.3%), and NABirds (88.1%).

17.3DSSD: Point-based 3D Single Stage Object Detector ⬇️

Currently, there have been many kinds of voxel-based 3D single stage detectors, while point-based single stage methods are still underexplored. In this paper, we first present a lightweight and effective point-based 3D single stage object detector, named 3DSSD, achieving a good balance between accuracy and efficiency. In this paradigm, all upsampling layers and refinement stage, which are indispensable in all existing point-based methods, are abandoned to reduce the large computation cost. We novelly propose a fusion sampling strategy in downsampling process to make detection on less representative points feasible. A delicate box prediction network including a candidate generation layer, an anchor-free regression head with a 3D center-ness assignment strategy is designed to meet with our demand of accuracy and speed. Our paradigm is an elegant single stage anchor-free framework, showing great superiority to other existing methods. We evaluate 3DSSD on widely used KITTI dataset and more challenging nuScenes dataset. Our method outperforms all state-of-the-art voxel-based single stage methods by a large margin, and has comparable performance to two stage point-based methods as well, with inference speed more than 25 FPS, 2x faster than former state-of-the-art point-based methods.

18.HRank: Filter Pruning using High-Rank Feature Map ⬇️

Neural network pruning offers a promising prospect to facilitate deploying deep neural networks on resource-limited devices. However, existing methods are still challenged by the training inefficiency and labor cost in pruning designs, due to missing theoretical guidance of non-salient network components. In this paper, we propose a novel filter pruning method by exploring the High Rank of feature maps (HRank). Our HRank is inspired by the discovery that the average rank of multiple feature maps generated by a single filter is always the same, regardless of the number of image batches CNNs receive. Based on HRank, we develop a method that is mathematically formulated to prune filters with low-rank feature maps. The principle behind our pruning is that low-rank feature maps contain less information, and thus pruned results can be easily reproduced. Besides, we experimentally show that weights with high-rank feature maps contain more important information, such that even when a portion is not updated, very little damage would be done to the model performance. Without introducing any additional constraints, HRank leads to significant improvements over the state-of-the-arts in terms of FLOPs and parameters reduction, with similar accuracies. For example, with ResNet-110, we achieve a 58.2%-FLOPs reduction by removing 59.2% of the parameters, with only a small loss of 0.14% in top-1 accuracy on CIFAR-10. With Res-50, we achieve a 43.8%-FLOPs reduction by removing 36.7% of the parameters, with only a loss of 1.17% in the top-1 accuracy on ImageNet. The codes can be available at this https URL.

19.Improving STDP-based Visual Feature Learning with Whitening ⬇️

In recent years, spiking neural networks (SNNs) emerge as an alternative to deep neural networks (DNNs). SNNs present a higher computational efficiency using low-power neuromorphic hardware and require less labeled data for training using local and unsupervised learning rules such as spike timing-dependent plasticity (STDP). SNN have proven their effectiveness in image classification on simple datasets such as MNIST. However, to process natural images, a pre-processing step is required. Difference-of-Gaussians (DoG) filtering is typically used together with on-center/off-center coding, but it results in a loss of information that is detrimental to the classification performance. In this paper, we propose to use whitening as a pre-processing step before learning features with STDP. Experiments on CIFAR-10 show that whitening allows STDP to learn visual features that are closer to the ones learned with standard neural networks, with a significantly increased classification performance as compared to DoG filtering. We also propose an approximation of whitening as convolution kernels that is computationally cheaper to learn and more suited to be implemented on neuromorphic hardware. Experiments on CIFAR-10 show that it performs similarly to regular whitening. Cross-dataset experiments on CIFAR-10 and STL-10 also show that it is fairly stable across datasets, making it possible to learn a single whitening transformation to process different datasets.

20.When Relation Networks meet GANs: Relation GANs with Triplet Loss ⬇️

Though recent research has achieved remarkable progress in generating realistic images with generative adversarial networks (GANs), the lack of training stability is still a lingering concern of most GANs, especially on high-resolution inputs and complex datasets. Since the randomly generated distribution can hardly overlap with the real distribution, training GANs often suffers from the gradient vanishing problem. A number of approaches have been proposed to address this issue by constraining the discriminator's capabilities using empirical techniques, like weight clipping, gradient penalty, spectral normalization etc. In this paper, we provide a more principled approach as an alternative solution to this issue. Instead of training the discriminator to distinguish real and fake input samples, we investigate the relationship between paired samples by training the discriminator to separate paired samples from the same distribution and those from different distributions. To this end, we explore a relation network architecture for the discriminator and design a triplet loss which performs better generalization and stability. Extensive experiments on benchmark datasets show that the proposed relation discriminator and new loss can provide significant improvement on variable vision tasks including unconditional and conditional image generation and image translation. Our source codes are available on the website: \url{this https URL}

21.Audio-driven Talking Face Video Generation with Natural Head Pose ⬇️

Real-world talking faces often accompany with natural head movement. However, most existing talking face video generation methods only consider facial animation with fixed head pose. In this paper, we address this problem by proposing a deep neural network model that takes an audio signal A of a source person and a very short video V of a target person as input, and outputs a synthesized high-quality talking face video with natural head pose (making use of the visual information in V), expression and lip synchronization (by considering both A and V). The most challenging issue in our work is that natural poses often cause in-plane and out-of-plane head rotations, which makes synthesized talking face video far from realistic. To address this challenge, we reconstruct 3D face animation and re-render it into synthesized frames. To fine tune these frames into realistic ones with smooth background transition, we propose a novel memory-augmented GAN module. Extensive experiments and three user studies show that our method can generate high-quality (i.e., natural head movements, expressions and good lip synchronization) personalized talking face videos, outperforming the state-of-the-art methods.

22.Semantic Flow for Fast and Accurate Scene Parsing ⬇️

In this paper, we focus on effective methods for fast and accurate scene parsing. A common practice to improve the performance is to attain high resolution feature maps with strong semantic representation. Two strategies are widely used---astrous convolutions and feature pyramid fusion, are either computation intensive or ineffective. Inspired by Optical Flow for motion alignment between adjacent video frames, we propose a Flow Alignment Module (FAM) to learn Semantic Flow between feature maps of adjacent levels and broadcast high-level features to high resolution features effectively and efficiently. Furthermore, integrating our module to a common feature pyramid structure exhibits superior performance over other real-time methods even on very light-weight backbone networks, such as ResNet-18. Extensive experiments are conducted on several challenging datasets, including Cityscapes, PASCAL Context, ADE20K and CamVid. Particularly, our network is the first to achieve 80.4% mIoU on Cityscapes with a frame rate of 26 FPS. The code will be available at \url{this https URL}.

23.DeepSign: Deep On-Line Signature Verification ⬇️

Deep learning has become a breathtaking technology in the last years, overcoming traditional handcrafted approaches and even humans for many different tasks. However, in some tasks, such as the verification of handwritten signatures, the amount of publicly available data is scarce, what makes difficult to test the real limits of deep learning. In addition to the lack of public data, it is not easy to evaluate the improvements of novel proposed approaches as different databases and experimental protocols are usually considered.
The main contributions of this study are: i) we provide an in-depth analysis of state-of-the-art deep learning approaches for on-line signature verification, ii) we present and describe the new DeepSignDB on-line handwritten signature biometric public database, iii) we propose a standard experimental protocol and benchmark to be used for the research community in order to perform a fair comparison of novel approaches with the state of the art, and iv) we adapt and evaluate our recent deep learning approach named Time-Aligned Recurrent Neural Networks (TA-RNNs) for the task of on-line handwritten signature verification. This approach combines the potential of Dynamic Time Warping and Recurrent Neural Networks to train more robust systems against forgeries. Our proposed TA-RNN system outperforms the state of the art, achieving results even below 2.0% EER when considering skilled forgery impostors and just one training signature per user.

24.SMOKE: Single-Stage Monocular 3D Object Detection via Keypoint Estimation ⬇️

Estimating 3D orientation and translation of objects is essential for infrastructure-less autonomous navigation and driving. In case of monocular vision, successful methods have been mainly based on two ingredients: (i) a network generating 2D region proposals, (ii) a R-CNN structure predicting 3D object pose by utilizing the acquired regions of interest. We argue that the 2D detection network is redundant and introduces non-negligible noise for 3D detection. Hence, we propose a novel 3D object detection method, named SMOKE, in this paper that predicts a 3D bounding box for each detected object by combining a single keypoint estimate with regressed 3D variables. As a second contribution, we propose a multi-step disentangling approach for constructing the 3D bounding box, which significantly improves both training convergence and detection accuracy. In contrast to previous 3D detection techniques, our method does not require complicated pre/post-processing, extra data, and a refinement stage. Despite of its structural simplicity, our proposed SMOKE network outperforms all existing monocular 3D detection methods on the KITTI dataset, giving the best state-of-the-art result on both 3D object detection and Bird's eye view evaluation. The code will be made publicly available.

25.GANHopper: Multi-Hop GAN for Unsupervised Image-to-Image Translation ⬇️

We introduce GANHOPPER, an unsupervised image-to-image translation network that transforms images gradually between two domains, through multiple hops. Instead of executing translation directly, we steer the translation by requiring the network to produce in-between images which resemble weighted hybrids between images from the two in-put domains. Our network is trained on unpaired images from the two domains only, without any in-between images.All hops are produced using a single generator along each direction. In addition to the standard cycle-consistency and adversarial losses, we introduce a new hybrid discrimina-tor, which is trained to classify the intermediate images produced by the generator as weighted hybrids, with weights based on a predetermined hop count. We also introduce a smoothness term to constrain the magnitude of each hop,further regularizing the translation. Compared to previous methods, GANHOPPER excels at image translations involving domain-specific image features and geometric variations while also preserving non-domain-specific features such as backgrounds and general color schemes.

26.LeafGAN: An Effective Data Augmentation Method for Practical Plant Disease Diagnosis ⬇️

Many applications for the automated diagnosis of plant disease have been developed based on the success of deep learning techniques. However, these applications often suffer from overfitting, and the diagnostic performance is drastically decreased when used on test datasets from new environments. The typical reasons for this are that the symptoms to be detected are unclear or faint, and there are limitations related to data diversity. In this paper, we propose LeafGAN, a novel image-to-image translation system with own attention mechanism. LeafGAN generates a wide variety of diseased images via transformation from healthy images, as a data augmentation tool for improving the performance of plant disease diagnosis. Thanks to its own attention mechanism, our model can transform only relevant areas from images with a variety of backgrounds, thus enriching the versatility of the training images. Experiments with five-class cucumber disease classification show that data augmentation with vanilla CycleGAN cannot help to improve the generalization, i.e. disease diagnostic performance increased by only 0.7% from the baseline. In contrast, LeafGAN boosted the diagnostic performance by 7.4%. We also visually confirmed the generated images by our LeafGAN were much better quality and more convincing than those generated by vanilla CycleGAN.

27.Utilizing a null class to restrict decision spaces and defend against neural network adversarial attacks ⬇️

Despite recent progress, deep neural networks generally continue to be vulnerable to so-called adversarial examples--input images with small perturbations that can result in changes in the output classifications, despite no such change in the semantic meaning to human viewers. This is true even for seemingly simple challenges such as the MNIST digit classification task. In part, this suggests that these networks are not relying on the same set of object features as humans use to make these classifications. In this paper we examine an additional, and largely unexplored, cause behind this phenomenon--namely, the use of the conventional training paradigm in which the entire input space is parcellated among the training classes. Owing to this paradigm, learned decision spaces for individual classes span excessively large regions of the input space and include images that have no semantic similarity to images in the training set. In this study, we train models that include a null class. That is, models may "opt-out" of classifying an input image as one of the digit classes. During training, null images are created through a variety of methods, in an attempt to create tighter and more semantically meaningful decision spaces for the digit classes. The best performing models classify nearly all adversarial examples as nulls, rather than mistaking them as a member of an incorrect digit class, while simultaneously maintaining high accuracy on the unperturbed test set. The use of a null class and the training paradigm presented herein may provide an effective defense against adversarial attacks for some applications. Code for replicating this study will be made available at this https URL .

28.Triple Wins: Boosting Accuracy, Robustness and Efficiency Together by Enabling Input-Adaptive Inference ⬇️

Deep networks were recently suggested to face the odds between accuracy (on clean natural images) and robustness (on adversarially perturbed images) (Tsipras et al., 2019). Such a dilemma is shown to be rooted in the inherently higher sample complexity (Schmidt et al., 2018) and/or model capacity (Nakkiran, 2019), for learning a high-accuracy and robust classifier. In view of that, give a classification task, growing the model capacity appears to help draw a win-win between accuracy and robustness, yet at the expense of model size and latency, therefore posing challenges for resource-constrained applications. Is it possible to co-design model accuracy, robustness and efficiency to achieve their triple wins? This paper studies multi-exit networks associated with input-adaptive efficient inference, showing their strong promise in achieving a "sweet point" in cooptimizing model accuracy, robustness and efficiency. Our proposed solution, dubbed Robust Dynamic Inference Networks (RDI-Nets), allows for each input (either clean or adversarial) to adaptively choose one of the multiple output layers (early branches or the final one) to output its prediction. That multi-loss adaptivity adds new variations and flexibility to adversarial attacks and defenses, on which we present a systematical investigation. We show experimentally that by equipping existing backbones with such robust adaptive inference, the resulting RDI-Nets can achieve better accuracy and robustness, yet with over 30% computational savings, compared to the defended original models.

29.Multi-Stream Networks and Ground-Truth Generation for Crowd Counting ⬇️

Crowd scene analysis has received a lot of attention recently due to the wide variety of applications, for instance, forensic science, urban planning, surveillance and security. In this context, a challenging task is known as crowd counting, whose main purpose is to estimate the number of people present in a single image. A Multi-Stream Convolutional Neural Network is developed and evaluated in this work, which receives an image as input and produces a density map that represents the spatial distribution of people in an end-to-end fashion. In order to address complex crowd counting issues, such as extremely unconstrained scale and perspective changes, the network architecture utilizes receptive fields with different size filters for each stream. In addition, we investigate the influence of the two most common fashions on the generation of ground truths and propose a hybrid method based on tiny face detection and scale interpolation. Experiments conducted on two challenging datasets, UCF-CC-50 and ShanghaiTech, demonstrate that using our ground truth generation methods achieves superior results.

30.Exploring Spatial-Temporal Multi-Frequency Analysis for High-Fidelity and Temporal-Consistency Video Prediction ⬇️

Video prediction is a pixel-wise dense prediction task to infer future frames based on past frames. Missing appearance details and motion blur are still two major problems for current predictive models, which lead to image distortion and temporal inconsistency. In this paper, we point out the necessity of exploring multi-frequency analysis to deal with the two problems. Inspired by the frequency band decomposition characteristic of Human Vision System (HVS), we propose a video prediction network based on multi-level wavelet analysis to deal with spatial and temporal information in a unified manner. Specifically, the multi-level spatial discrete wavelet transform decomposes each video frame into anisotropic sub-bands with multiple frequencies, helping to enrich structural information and reserve fine details. On the other hand, multi-level temporal discrete wavelet transform which operates on time axis decomposes the frame sequence into sub-band groups of different frequencies to accurately capture multi-frequency motions under a fixed frame rate. Extensive experiments on diverse datasets demonstrate that our model shows significant improvements on fidelity and temporal consistency over state-of-the-art works.

31.DotFAN: A Domain-transferred Face Augmentation Network for Pose and Illumination Invariant Face Recognition ⬇️

The performance of a convolutional neural network (CNN) based face recognition model largely relies on the richness of labelled training data. Collecting a training set with large variations of a face identity under different poses and illumination changes, however, is very expensive, making the diversity of within-class face images a critical issue in practice. In this paper, we propose a 3D model-assisted domain-transferred face augmentation network (DotFAN) that can generate a series of variants of an input face based on the knowledge distilled from existing rich face datasets collected from other domains. DotFAN is structurally a conditional CycleGAN but has two additional subnetworks, namely face expert network (FEM) and face shape regressor (FSR), for latent code control. While FSR aims to extract face attributes, FEM is designed to capture a face identity. With their aid, DotFAN can learn a disentangled face representation and effectively generate face images of various facial attributes while preserving the identity of augmented faces. Experiments show that DotFAN is beneficial for augmenting small face datasets to improve their within-class diversity so that a better face recognition model can be learned from the augmented dataset.

32.Assembling Semantically-Disentangled Representations for Predictive-Generative Models via Adaptation from Synthetic Domain ⬇️

Deep neural networks can form high-level hierarchical representations of input data. Various researchers have demonstrated that these representations can be used to enable a variety of useful applications. However, such representations are typically based on the statistics within the data, and may not conform with the semantic representation that may be necessitated by the application. Conditional models are typically used to overcome this challenge, but they require large annotated datasets which are difficult to come by and costly to create. In this paper, we show that semantically-aligned representations can be generated instead with the help of a physics based engine. This is accomplished by creating a synthetic dataset with decoupled attributes, learning an encoder for the synthetic dataset, and augmenting prescribed attributes from the synthetic domain with attributes from the real domain. It is shown that the proposed (SYNTH-VAE-GAN) method can construct a conditional predictive-generative model of human face attributes without relying on real data labels.

33.Random Bundle: Brain Metastases Segmentation Ensembling through Annotation Randomization ⬇️

We introduce a novel ensembling method, Random Bundle (RB), that improves performance for brain metastases segmentation. We create our ensemble by training each network on our dataset with 50% of our annotated lesions censored out. We also apply a lopsided bootstrap loss to recover performance after inducing an in silico 50% false negative rate and make our networks more sensitive. We improve our network detection of lesions's mAP value by 39% and more than triple the sensitivity at 80% precision. We also show slight improvements in segmentation quality through DICE score. Further, RB ensembling improves performance over baseline by a larger margin than a variety of popular ensembling strategies. Finally, we show that RB ensembling is computationally efficient by comparing its performance to a single network when both systems are constrained to have the same compute.

34.Reliable Fidelity and Diversity Metrics for Generative Models ⬇️

Devising indicative evaluation metrics for the image generation task remains an open problem. The most widely used metric for measuring the similarity between real and generated images has been the Fréchet Inception Distance (FID) score. Because it does not differentiate the fidelity and diversity aspects of the generated images, recent papers have introduced variants of precision and recall metrics to diagnose those properties separately. In this paper, we show that even the latest version of the precision and recall metrics are not reliable yet. For example, they fail to detect the match between two identical distributions, they are not robust against outliers, and the evaluation hyperparameters are selected arbitrarily. We propose density and coverage metrics that solve the above issues. We analytically and experimentally show that density and coverage provide more interpretable and reliable signals for practitioners than the existing metrics. Code: this https URL.

35.VisionGuard: Runtime Detection of Adversarial Inputs to Perception Systems ⬇️

Deep neural network (DNN) models have proven to be vulnerable to adversarial attacks. In this paper, we propose VisionGuard, a novel attack- and dataset-agnostic and computationally-light defense mechanism for adversarial inputs to DNN-based perception systems. In particular, VisionGuard relies on the observation that adversarial images are sensitive to lossy compression transformations. Specifically, to determine if an image is adversarial, VisionGuard checks if the output of the target classifier on a given input image changes significantly after feeding it a transformed version of the image under investigation. Moreover, we show that VisionGuard is computationally-light both at runtime and design-time which makes it suitable for real-time applications that may also involve large-scale image domains. To highlight this, we demonstrate the efficiency of VisionGuard on ImageNet, a task that is computationally challenging for the majority of relevant defenses. Finally, we include extensive comparative experiments on the MNIST, CIFAR10, and ImageNet datasets that show that VisionGuard outperforms existing defenses in terms of scalability and detection performance.

36.Shallow2Deep: Indoor Scene Modeling by Single Image Understanding ⬇️

Dense indoor scene modeling from 2D images has been bottlenecked due to the absence of depth information and cluttered occlusions. We present an automatic indoor scene modeling approach using deep features from neural networks. Given a single RGB image, our method simultaneously recovers semantic contents, 3D geometry and object relationship by reasoning indoor environment context. Particularly, we design a shallow-to-deep architecture on the basis of convolutional networks for semantic scene understanding and modeling. It involves multi-level convolutional networks to parse indoor semantics/geometry into non-relational and relational knowledge. Non-relational knowledge extracted from shallow-end networks (e.g. room layout, object geometry) is fed forward into deeper levels to parse relational semantics (e.g. support relationship). A Relation Network is proposed to infer the support relationship between objects. All the structured semantics and geometry above are assembled to guide a global optimization for 3D scene modeling. Qualitative and quantitative analysis demonstrates the feasibility of our method in understanding and modeling semantics-enriched indoor scenes by evaluating the performance of reconstruction accuracy, computation performance and scene complexity.

37.Robust Multimodal Brain Tumor Segmentation via Feature Disentanglement and Gated Fusion ⬇️

Accurate medical image segmentation commonly requires effective learning of the complementary information from multimodal data. However, in clinical practice, we often encounter the problem of missing imaging modalities. We tackle this challenge and propose a novel multimodal segmentation framework which is robust to the absence of imaging modalities. Our network uses feature disentanglement to decompose the input modalities into the modality-specific appearance code, which uniquely sticks to each modality, and the modality-invariant content code, which absorbs multimodal information for the segmentation task. With enhanced modality-invariance, the disentangled content code from each modality is fused into a shared representation which gains robustness to missing data. The fusion is achieved via a learning-based strategy to gate the contribution of different modalities at different locations. We validate our method on the important yet challenging multimodal brain tumor segmentation task with the BRATS challenge dataset. With competitive performance to the state-of-the-art approaches for full modality, our method achieves outstanding robustness under various missing modality(ies) situations, significantly exceeding the state-of-the-art method by over 16% in average for Dice on whole tumor segmentation.

38.Temporal Sparse Adversarial Attack on Gait Recognition ⬇️

Gait recognition has a broad application in social security due to its advantages in long-distance human identification. Despite the high accuracy of gait recognition systems, their adversarial robustness has not been explored. In this paper, we demonstrate that the state-of-the-art gait recognition model is vulnerable to adversarial attacks. A novel temporal sparse adversarial attack under a new defined distortion measurement is proposed. GAN-based architecture is employed to semantically generate adversarial high-quality gait silhouette. By sparsely substituting or inserting a few adversarial gait silhouettes, our proposed method can achieve a high attack success rate. The imperceptibility and the attacking success rate of the adversarial examples are well balanced. Experimental results show even only one-fortieth frames are attacked, the attack success rate still reaches 76.8%.

39.Active Lighting Recurrence by Parallel Lighting Analogy for Fine-Grained Change Detection ⬇️

This paper studies a new problem, namely active lighting recurrence (ALR) that physically relocalizes a light source to reproduce the lighting condition from single reference image for a same scene, which may suffer from fine-grained changes during twice observations. ALR is of great importance for fine-grained visual inspection and change detection, because some phenomena or minute changes can only be clearly observed under particular lighting conditions. Therefore, effective ALR should be able to online navigate a light source toward the target pose, which is challenging due to the complexity and diversity of real-world lighting and imaging processes. To this end, we propose to use the simple parallel lighting as an analogy model and based on Lambertian law to compose an instant navigation ball for this purpose. We theoretically prove the feasibility, i.e., equivalence and convergence, of this ALR approach for realistic near point light source and small near surface light source. Besides, we also theoretically prove the invariance of our ALR approach to the ambiguity of normal and lighting decomposition. The effectiveness and superiority of the proposed approach have been verified by both extensive quantitative experiments and challenging real-world tasks on fine-grained change detection of cultural heritages. We also validate the generality of our approach to non-Lambertian scenes.

40.UnMask: Adversarial Detection and Defense Through Robust Feature Alignment ⬇️

Deep learning models are being integrated into a wide range of high-impact, security-critical systems, from self-driving cars to medical diagnosis. However, recent research has demonstrated that many of these deep learning architectures are vulnerable to adversarial attacks--highlighting the vital need for defensive techniques to detect and mitigate these attacks before they occur. To combat these adversarial attacks, we developed UnMask, an adversarial detection and defense framework based on robust feature alignment. The core idea behind UnMask is to protect these models by verifying that an image's predicted class ("bird") contains the expected robust features (e.g., beak, wings, eyes). For example, if an image is classified as "bird", but the extracted features are wheel, saddle and frame, the model may be under attack. UnMask detects such attacks and defends the model by rectifying the misclassification, re-classifying the image based on its robust features. Our extensive evaluation shows that UnMask (1) detects up to 96.75% of attacks, with a false positive rate of 9.66% and (2) defends the model by correctly classifying up to 93% of adversarial images produced by the current strongest attack, Projected Gradient Descent, in the gray-box setting. UnMask provides significantly better protection than adversarial training across 8 attack vectors, averaging 31.18% higher accuracy. Our proposed method is architecture agnostic and fast. We open source the code repository and data with this paper: this https URL.

41.Particle Filter Based Monocular Human Tracking with a 3D Cardbox Model and a Novel Deterministic Resampling Strategy ⬇️

The challenge of markerless human motion tracking is the high dimensionality of the search space. Thus, efficient exploration in the search space is of great significance. In this paper, a motion capturing algorithm is proposed for upper body motion tracking. The proposed system tracks human motion based on monocular silhouette-matching, and it is built on the top of a hierarchical particle filter, within which a novel deterministic resampling strategy (DRS) is applied. The proposed system is evaluated quantitatively with the ground truth data measured by an inertial sensor system. In addition, we compare the DRS with the stratified resampling strategy (SRS). It is shown in experiments that DRS outperforms SRS with the same amount of particles. Moreover, a new 3D articulated human upper body model with the name 3D cardbox model is created and is proven to work successfully for motion tracking. Experiments show that the proposed system can robustly track upper body motion without self-occlusion. Motions towards the camera can also be well tracked.

42.Image to Language Understanding: Captioning approach ⬇️

Extracting context from visual representations is of utmost importance in the advancement of Computer Science. Representation of such a format in Natural Language has a huge variety of applications such as helping the visually impaired etc. Such an approach is a combination of Computer Vision and Natural Language techniques which is a hard problem to solve. This project aims to compare different approaches for solving the image captioning problem. In specific, the focus was on comparing two different types of models: Encoder-Decoder approach and a Multi-model approach. In the encoder-decoder approach, inject and merge architectures were compared against a multi-modal image captioning approach based primarily on object detection. These approaches have been compared on the basis on state of the art sentence comparison metrics such as BLEU, GLEU, Meteor, and Rouge on a subset of the Google Conceptual captions dataset which contains 100k images. On the basis of this comparison, we observed that the best model was the Inception injected encoder model. This best approach has been deployed as a web-based system. On uploading an image, such a system will output the best caption associated with the image.

43.Kullback-Leibler Divergence-Based Fuzzy $C$-Means Clustering Incorporating Morphological Reconstruction and Wavelet Frames for Image Segmentation ⬇️

Although spatial information of images usually enhance the robustness of the Fuzzy C-Means (FCM) algorithm, it greatly increases the computational costs for image segmentation. To achieve a sound trade-off between the segmentation performance and the speed of clustering, we come up with a Kullback-Leibler (KL) divergence-based FCM algorithm by incorporating a tight wavelet frame transform and a morphological reconstruction operation. To enhance FCM's robustness, an observed image is first filtered by using the morphological reconstruction. A tight wavelet frame system is employed to decompose the observed and filtered images so as to form their feature sets. Considering these feature sets as data of clustering, an modified FCM algorithm is proposed, which introduces a KL divergence term in the partition matrix into its objective function. The KL divergence term aims to make membership degrees of each image pixel closer to those of its neighbors, which brings that the membership partition becomes more suitable and the parameter setting of FCM becomes simplified. On the basis of the obtained partition matrix and prototypes, the segmented feature set is reconstructed by minimizing the inverse process of the modified objective function. To modify abnormal features produced in the reconstruction process, each reconstructed feature is reassigned to the closest prototype. As a result, the segmentation accuracy of KL divergence-based FCM is further improved. What's more, the segmented image is reconstructed by using a tight wavelet frame reconstruction operation. Finally, supporting experiments coping with synthetic, medical and color images are reported. Experimental results exhibit that the proposed algorithm works well and comes with better segmentation performance than other comparative algorithms. Moreover, the proposed algorithm requires less time than most of the FCM-related algorithms.

44.Deep Nearest Neighbor Anomaly Detection ⬇️

Nearest neighbors is a successful and long-standing technique for anomaly detection. Significant progress has been recently achieved by self-supervised deep methods (e.g. RotNet). Self-supervised features however typically under-perform Imagenet pre-trained features. In this work, we investigate whether the recent progress can indeed outperform nearest-neighbor methods operating on an Imagenet pretrained feature space. The simple nearest-neighbor based-approach is experimentally shown to outperform self-supervised methods in: accuracy, few shot generalization, training time and noise robustness while making fewer assumptions on image distributions.

45.Batch Normalization Biases Deep Residual Networks Towards Shallow Paths ⬇️

Batch normalization has multiple benefits. It improves the conditioning of the loss landscape, and is a surprisingly effective regularizer. However, the most important benefit of batch normalization arises in residual networks, where it dramatically increases the largest trainable depth. We identify the origin of this benefit: At initialization, batch normalization downscales the residual branch relative to the skip connection, by a normalizing factor proportional to the square root of the network depth. This ensures that, early in training, the function computed by deep normalized residual networks is dominated by shallow paths with well-behaved gradients. We use this insight to develop a simple initialization scheme which can train very deep residual networks without normalization. We also clarify that, although batch normalization does enable stable training with larger learning rates, this benefit is only useful when one wishes to parallelize training over large batch sizes. Our results help isolate the distinct benefits of batch normalization in different architectures.

46.Group Membership Verification with Privacy: Sparse or Dense? ⬇️

Group membership verification checks if a biometric trait corresponds to one member of a group without revealing the identity of that member. Recent contributions provide privacy for group membership protocols through the joint use of two mechanisms: quantizing templates into discrete embeddings and aggregating several templates into one group representation. However, this scheme has one drawback: the data structure representing the group has a limited size and cannot recognize noisy queries when many templates are aggregated. Moreover, the sparsity of the embeddings seemingly plays a crucial role on the performance verification. This paper proposes a mathematical model for group membership verification allowing to reveal the impact of sparsity on both security, compactness, and verification performances. This model bridges the gap towards a Bloom filter robust to noisy queries. It shows that a dense solution is more competitive unless the queries are almost noiseless.

47.Self-Adaptive Training: beyond Empirical Risk Minimization ⬇️

We propose self-adaptive training---a new training algorithm that dynamically corrects problematic training labels by model predictions without incurring extra computational cost---to improve generalization of deep learning for potentially corrupted training data. This problem is crucial towards robustly learning from data that are corrupted by, e.g., label noises and out-of-distribution samples. The standard empirical risk minimization (ERM) for such data, however, may easily overfit noises and thus suffers from sub-optimal performance. In this paper, we observe that model predictions can substantially benefit the training process: self-adaptive training significantly improves generalization over ERM under various levels of noises, and mitigates the overfitting issue in both natural and adversarial training. We evaluate the error-capacity curve of self-adaptive training: the test error is monotonously decreasing w.r.t. model capacity. This is in sharp contrast to the recently-discovered double-descent phenomenon in ERM which might be a result of overfitting of noises. Experiments on CIFAR and ImageNet datasets verify the effectiveness of our approach in two applications: classification with label noise and selective classification. We release our code at \url{this https URL}.

48.TensorShield: Tensor-based Defense Against Adversarial Attacks on Images ⬇️

Recent studies have demonstrated that machine learning approaches like deep neural networks (DNNs) are easily fooled by adversarial attacks. Subtle and imperceptible perturbations of the data are able to change the result of deep neural networks. Leveraging vulnerable machine learning methods raises many concerns especially in domains where security is an important factor. Therefore, it is crucial to design defense mechanisms against adversarial attacks. For the task of image classification, unnoticeable perturbations mostly occur in the high-frequency spectrum of the image. In this paper, we utilize tensor decomposition techniques as a preprocessing step to find a low-rank approximation of images which can significantly discard high-frequency perturbations. Recently a defense framework called Shield could "vaccinate" Convolutional Neural Networks (CNN) against adversarial examples by performing random-quality JPEG compressions on local patches of images on the ImageNet dataset. Our tensor-based defense mechanism outperforms the SLQ method from Shield by 14% against FastGradient Descent (FGSM) adversarial attacks, while maintaining comparable speed.

49.Beyond Camera Motion Removing: How to Handle Outliers in Deblurring ⬇️

Performing camera motion deblurring is an important low-level vision task for achieving better imaging quality. When a scene has outliers such as saturated pixels and salt-and pepper noise, the image becomes more difficult to restore. In this paper, we propose an edge-aware scalerecurrent network (EASRN) to conduct camera motion deblurring. EASRN has a separate deblurring module that removes blur at multiple scales and an upsampling module that fuses different input scales. We propose a salient edge detection network to supervise the training process and solve the outlier problem by proposing a novel method of dataset generation. Light streaks are printed on the sharp image to simulate the cutoff effect from saturation. We evaluate our method on the standard deblurring datasets. Both objective evaluation indexes and subjective visualization show that our method results in better deblurring quality than the other state-of-the-art approaches.

50.Real-time Kinematic Ground Truth for the Oxford RobotCar Dataset ⬇️

We describe the release of reference data towards a challenging long-term localisation and mapping benchmark based on the large-scale Oxford RobotCar Dataset. The release includes 72 traversals of a route through Oxford, UK, gathered in all illumination, weather and traffic conditions, and is representative of the conditions an autonomous vehicle would be expected to operate reliably in. Using post-processed raw GPS, IMU, and static GNSS base station recordings, we have produced a globally-consistent centimetre-accurate ground truth for the entire year-long duration of the dataset. Coupled with a planned online benchmarking service, we hope to enable quantitative evaluation and comparison of different localisation and mapping approaches focusing on long-term autonomy for road vehicles in urban environments challenged by changing weather.

51.Communication Contention Aware Scheduling of Multiple Deep Learning Training Jobs ⬇️

Distributed Deep Learning (DDL) has rapidly grown its popularity since it helps boost the training performance on high-performance GPU clusters. Efficient job scheduling is indispensable to maximize the overall performance of the cluster when training multiple jobs simultaneously. However, existing schedulers do not consider the communication contention of multiple communication tasks from different distributed training jobs, which could deteriorate the system performance and prolong the job completion time. In this paper, we first establish a new DDL job scheduling framework which organizes DDL jobs as Directed Acyclic Graphs (DAGs) and considers communication contention between nodes. We then propose an efficient algorithm, LWF-$\kappa$, to balance the GPU utilization and consolidate the allocated GPUs for each job. When scheduling those communication tasks, we observe that neither avoiding all the contention nor blindly accepting them is optimal to minimize the job completion time. We thus propose a provable algorithm, AdaDUAL, to efficiently schedule those communication tasks. Based on AdaDUAL, we finally propose Ada-SRSF for the DDL job scheduling problem. Simulations on a 64-GPU cluster connected with 10 Gbps Ethernet show that LWF-$\kappa$ achieves up to $1.59\times$ improvement over the classical first-fit algorithms. More importantly, Ada-SRSF reduces the average job completion time by $20.1%$ and $36.7%$, as compared to the SRSF(1) scheme (avoiding all the contention) and the SRSF(2) scheme (blindly accepting all of two-way communication contention) respectively.

52.Implicit Geometric Regularization for Learning Shapes ⬇️

Representing shapes as level sets of neural networks has been recently proved to be useful for different shape analysis and reconstruction tasks So far, such representations were computed using either: (i) pre-computed implicit shape representations; or (ii) loss functions explicitly defined over the neural level sets. In this paper we offer a new paradigm for computing high fidelity implicit neural representations directly from raw data (i.e., point clouds, with or without normal information). We observe that a rather simple loss function, encouraging the neural network to vanish on the input point cloud and to have a unit norm gradient, possesses an implicit geometric regularization property that favors smooth and natural zero level set surfaces, avoiding bad zero-loss solutions. We provide a theoretical analysis of this property for the linear case, and show that, in practice, our method leads to state of the art implicit neural representations with higher level-of-details and fidelity compared to previous methods.

53.Generalized Octave Convolutions for Learned Multi-Frequency Image Compression ⬇️

Learned image compression has recently shown the potential to outperform all standard codecs. The state-of-the-art rate-distortion performance has been achieved by context-adaptive entropy approaches in which hyperprior and autoregressive models are jointly utilized to effectively capture the spatial dependencies in the latent representations. However, the latents contain a mixture of high and low frequency information, which has inefficiently been represented by features maps of the same spatial resolution in previous works. In this paper, we propose the first learned multi-frequency image compression approach that uses the recently developed octave convolutions to factorize the latents into high and low frequencies. Since the low frequency is represented by a lower resolution, their spatial redundancy is reduced, which improves the compression rate. Moreover, octave convolutions impose effective high and low frequency communication, which can improve the reconstruction quality. We also develop novel generalized octave convolution and octave transposed-convolution architectures with internal activation layers to preserve the spatial structure of the information. Our experiments show that the proposed scheme outperforms all standard codecs and learning-based methods in both PSNR and MS-SSIM metrics, and establishes the new state of the art for learned image compression.

54.Deep Multimodal Image-Text Embeddings for Automatic Cross-Media Retrieval ⬇️

This paper considers the task of matching images and sentences by learning a visual-textual embedding space for cross-modal retrieval. Finding such a space is a challenging task since the features and representations of text and image are not comparable. In this work, we introduce an end-to-end deep multimodal convolutional-recurrent network for learning both vision and language representations simultaneously to infer image-text similarity. The model learns which pairs are a match (positive) and which ones are a mismatch (negative) using a hinge-based triplet ranking. To learn about the joint representations, we leverage our newly extracted collection of tweets from Twitter. The main characteristic of our dataset is that the images and tweets are not standardized the same as the benchmarks. Furthermore, there can be a higher semantic correlation between the pictures and tweets contrary to benchmarks in which the descriptions are well-organized. Experimental results on MS-COCO benchmark dataset show that our model outperforms certain methods presented previously and has competitive performance compared to the state-of-the-art. The code and dataset have been made available publicly.

55.NeurIPS 2019 Disentanglement Challenge: Improved Disentanglement through Aggregated Convolutional Feature Maps ⬇️

This report to our stage 1 submission to the NeurIPS 2019 disentanglement challenge presents a simple image preprocessing method for training VAEs leading to improved disentanglement compared to directly using the images. In particular, we propose to use regionally aggregated feature maps extracted from CNNs pretrained on ImageNet. Our method achieved the 2nd place in stage 1 of the challenge. Code is available at this https URL.

56.Gradual Channel Pruning while Training using Feature Relevance Scores for Convolutional Neural Networks ⬇️

The enormous inference cost of deep neural networks can be scaled down by network compression. Pruning is one of the predominant approaches used for deep network compression. However, existing pruning techniques have one or more of the following limitations: 1) Additional energy cost on top of the compute heavy training stage due to pruning and fine-tuning stages, 2) Layer-wise pruning based on the statistics of a particular, ignoring the effect of error propagation in the network, 3) Lack of an efficient estimate for determining the important channels globally, 4) Unstructured pruning requires specialized hardware for effective use. To address all the above issues, we present a simple-yet-effective gradual channel pruning while training methodology using a novel data driven metric referred as Feature relevance score. The proposed technique gets rid of the additional retraining cycles by pruning least important channels in a structured fashion at fixed intervals during the actual training phase. Feature relevance scores help in efficiently evaluating the contribution of each channel towards the discriminative power of the network. We demonstrate the effectiveness of the proposed methodology on architectures such as VGG and ResNet using datasets such as CIFAR-10, CIFAR-100 and ImageNet, and successfully achieve significant model compression while trading off less than $1%$ accuracy. Notably on CIFAR-10 dataset trained on ResNet-110, our approach achieves $2.4\times$ compression and a $56%$ reduction in FLOPs with an accuracy drop of $0.01%$ compared to the unpruned network.

57.Monocular Direct Sparse Localization in a Prior 3D Surfel Map ⬇️

In this paper, we introduce an approach to tracking the pose of a monocular camera in a prior surfel map. By rendering vertex and normal maps from the prior surfel map, the global planar information for the sparse tracked points in the image frame is obtained. The tracked points with and without the global planar information involve both global and local constraints of frames to the system. Our approach formulates all constraints in the form of direct photometric errors within a local window of the frames. The final optimization utilizes these constraints to provide the accurate estimation of global 6-DoF camera poses with the absolute scale. The extensive simulation and real-world experiments demonstrate that our monocular method can provide accurate camera localization results under various conditions.

58.Unsupervised Denoising for Satellite Imagery using Wavelet Subband CycleGAN ⬇️

Multi-spectral satellite imaging sensors acquire various spectral band images such as red (R), green (G), blue (B), near-infrared (N), etc. Thanks to the unique spectroscopic property of each spectral band with respective to the objects on the ground, multi-spectral satellite imagery can be used for various geological survey applications. Unfortunately, image artifacts from imaging sensor noises often affect the quality of scenes and have negative impacts on the applications of satellite imagery. Recently, deep learning approaches have been extensively explored for the removal of noises in satellite imagery. Most deep learning denoising methods, however, follow a supervised learning scheme, which requires matched noisy image and clean image pairs that are difficult to collect in real situations. In this paper, we propose a novel unsupervised multispectral denoising method for satellite imagery using wavelet subband cycle-consistent adversarial network (WavCycleGAN). The proposed method is based on unsupervised learning scheme using adversarial loss and cycle-consistency loss to overcome the lack of paired data. Moreover, in contrast to the standard image domain cycleGAN, we introduce a wavelet subband domain learning scheme for effective denoising without sacrificing high frequency components such as edges and detail information. Experimental results for the removal of vertical stripe and wave noises in satellite imaging sensors demonstrate that the proposed method effectively removes noises and preserves important high frequency features of satellite images.

59.Practical and Bilateral Privacy-preserving Federated Learning ⬇️

Federated learning, as an emerging distributed training model of neural networks without collecting raw data, has attracted widespread attention. However, almost all existing researches of federated learning only consider protecting the privacy of clients, but not preventing model iterates and final model parameters from leaking to untrusted clients and external attackers. In this paper, we present the first bilateral privacy-preserving federated learning scheme, which protects not only the raw training data of clients, but also model iterates during the training phase as well as final model parameters. Specifically, we present an efficient privacy-preserving technique to mask or encrypt the global model, which not only allows clients to train over the noisy global model, but also ensures only the server can obtain the exact updated model. Detailed security analysis shows that clients can access neither model iterates nor the final global model; meanwhile, the server cannot obtain raw training data of clients from additional information used for recovering the exact updated model. Finally, extensive experiments demonstrate the proposed scheme has comparable model accuracy with traditional federated learning without bringing much extra communication overhead.

60.Neuron Shapley: Discovering the Responsible Neurons ⬇️

We develop Neuron Shapley as a new framework to quantify the contribution of individual neurons to the prediction and performance of a deep network. By accounting for interactions across neurons, Neuron Shapley is more effective in identifying important filters compared to common approaches based on activation patterns. Interestingly, removing just 30 filters with the highest Shapley scores effectively destroys the prediction accuracy of Inception-v3 on ImageNet. Visualization of these few critical filters provides insights into how the network functions. Neuron Shapley is a flexible framework and can be applied to identify responsible neurons in many tasks. We illustrate additional applications of identifying filters that are responsible for biased prediction in facial recognition and filters that are vulnerable to adversarial attacks. Removing these filters is a quick way to repair models. Enabling all these applications is a new multi-arm bandit algorithm that we developed to efficiently estimate Neuron Shapley values.

61.HarDNN: Feature Map Vulnerability Evaluation in CNNs ⬇️

As Convolutional Neural Networks (CNNs) are increasingly being employed in safety-critical applications, it is important that they behave reliably in the face of hardware errors. Transient hardware errors may percolate undesirable state during execution, resulting in software-manifested errors which can adversely affect high-level decision making. This paper presents HarDNN, a software-directed approach to identify vulnerable computations during a CNN inference and selectively protect them based on their propensity towards corrupting the inference output in the presence of a hardware error. We show that HarDNN can accurately estimate relative vulnerability of a feature map (fmap) in CNNs using a statistical error injection campaign, and explore heuristics for fast vulnerability assessment. Based on these results, we analyze the tradeoff between error coverage and computational overhead that the system designers can use to employ selective protection. Results show that the improvement in resilience for the added computation is superlinear with HarDNN. For example, HarDNN improves SqueezeNet's resilience by 10x with just 30% additional computations.

62.Automatic Data Augmentation via Deep Reinforcement Learning for Effective Kidney Tumor Segmentation ⬇️

Conventional data augmentation realized by performing simple pre-processing operations (\eg, rotation, crop, \etc) has been validated for its advantage in enhancing the performance for medical image segmentation. However, the data generated by these conventional augmentation methods are random and sometimes harmful to the subsequent segmentation. In this paper, we developed a novel automatic learning-based data augmentation method for medical image segmentation which models the augmentation task as a trial-and-error procedure using deep reinforcement learning (DRL). In our method, we innovatively combine the data augmentation module and the subsequent segmentation module in an end-to-end training manner with a consistent loss. Specifically, the best sequential combination of different basic operations is automatically learned by directly maximizing the performance improvement (\ie, Dice ratio) on the available validation set. We extensively evaluated our method on CT kidney tumor segmentation which validated the promising results of our method.

63.Towards Label-Free 3D Segmentation of Optical Coherence Tomography Images of the Optic Nerve Head Using Deep Learning ⬇️

Since the introduction of optical coherence tomography (OCT), it has been possible to study the complex 3D morphological changes of the optic nerve head (ONH) tissues that occur along with the progression of glaucoma. Although several deep learning (DL) techniques have been recently proposed for the automated extraction (segmentation) and quantification of these morphological changes, the device specific nature and the difficulty in preparing manual segmentations (training data) limit their clinical adoption. With several new manufacturers and next-generation OCT devices entering the market, the complexity in deploying DL algorithms clinically is only increasing. To address this, we propose a DL based 3D segmentation framework that is easily translatable across OCT devices in a label-free manner (i.e. without the need to manually re-segment data for each device). Specifically, we developed 2 sets of DL networks. The first (referred to as the enhancer) was able to enhance OCT image quality from 3 OCT devices, and harmonized image-characteristics across these devices. The second performed 3D segmentation of 6 important ONH tissue layers. We found that the use of the enhancer was critical for our segmentation network to achieve device independency. In other words, our 3D segmentation network trained on any of 3 devices successfully segmented ONH tissue layers from the other two devices with high performance (Dice coefficients > 0.92). With such an approach, we could automatically segment images from new OCT devices without ever needing manual segmentation data from such devices.

64.Neural Architecture Search for Compressed Sensing Magnetic Resonance Image Reconstruction ⬇️

Recent works have demonstrated that deep learning (DL) based compressed sensing (CS) implementation can provide impressive improvements to reconstruct high-quality MR images from sub-sampling k-space data. However, network architectures adopted in current methods are all designed by handcraft, thus the performances of these networks are limited by researchers' expertise and labor. In this manuscript, we proposed a novel and efficient MR image reconstruction framework by Neural Architecture Search (NAS) algorithm. The inner cells in our reconstruction network are automatically defined from a flexible search space in a differentiable manner. Comparing to previous works where only several common convolutional operations are tried by human, our method can explore different operations (e.g. dilated convolution) with their possible combinations sufficiently. Our proposed method can also reach a better trade-off between computation cost and reconstruction performance for practical clinical translation. Experiments performed on a publicly available dataset show that our network produces better reconstruction results compared to the previous state-of-the-art methods in terms of PSNR and SSIM with 4 times fewer computation resources. The final network architecture found by the algorithm can also offer insights for network architecture designed in other medical image analysis applications.

65.Tuning-free Plug-and-Play Proximal Algorithm for Inverse Imaging Problems ⬇️

Plug-and-play (PnP) is a non-convex framework that combines ADMM or other proximal algorithms with advanced denoiser priors. Recently, PnP has achieved great empirical success, especially with the integration of deep learning-based denoisers. However, a key problem of PnP based approaches is that they require manual parameter tweaking. It is necessary to obtain high-quality results across the high discrepancy in terms of imaging conditions and varying scene content. In this work, we present a tuning-free PnP proximal algorithm, which can automatically determine the internal parameters including the penalty parameter, the denoising strength and the terminal time. A key part of our approach is to develop a policy network for automatic search of parameters, which can be effectively learned via mixed model-free and model-based deep reinforcement learning. We demonstrate, through numerical and visual experiments, that the learned policy can customize different parameters for different states, and often more efficient and effective than existing handcrafted criteria. Moreover, we discuss the practical considerations of the plugged denoisers, which together with our learned policy yield state-of-the-art results. This is prevalent on both linear and nonlinear exemplary inverse imaging problems, and in particular, we show promising results on Compressed Sensing MRI and phase retrieval.

66.Convex Shape Representation with Binary Labels for Image Segmentation: Models and Fast Algorithms ⬇️

We present a novel and effective binary representation for convex shapes. We show the equivalence between the shape convexity and some properties of the associated indicator function. The proposed method has two advantages. Firstly, the representation is based on a simple inequality constraint on the binary function rather than the definition of convex shapes, which allows us to obtain efficient algorithms for various applications with convexity prior. Secondly, this method is independent of the dimension of the concerned shape. In order to show the effectiveness of the proposed representation approach, we incorporate it with a probability based model for object segmentation with convexity prior. Efficient algorithms are given to solve the proposed models using Lagrange multiplier methods and linear approximations. Various experiments are given to show the superiority of the proposed methods.

67.Learning to Continually Learn ⬇️

Continual lifelong learning requires an agent or model to learn many sequentially ordered tasks, building on previous knowledge without catastrophically forgetting it. Much work has gone towards preventing the default tendency of machine learning models to catastrophically forget, yet virtually all such work involves manually-designed solutions to the problem. We instead advocate meta-learning a solution to catastrophic forgetting, allowing AI to learn to continually learn. Inspired by neuromodulatory processes in the brain, we propose A Neuromodulated Meta-Learning Algorithm (ANML). It differentiates through a sequential learning process to meta-learn an activation-gating function that enables context-dependent selective activation within a deep neural network. Specifically, a neuromodulatory (NM) neural network gates the forward pass of another (otherwise normal) neural network called the prediction learning network (PLN). The NM network also thus indirectly controls selective plasticity (i.e. the backward pass of) the PLN. ANML enables continual learning without catastrophic forgetting at scale: it produces state-of-the-art continual learning performance, sequentially learning as many as 600 classes (over 9,000 SGD updates).

68.Adversarial Attacks on Machine Learning Systems for High-Frequency Trading ⬇️

Algorithmic trading systems are often completely automated, and deep learning is increasingly receiving attention in this domain. Nonetheless, little is known about the robustness properties of these models. We study valuation models for algorithmic trading from the perspective of adversarial machine learning. We introduce new attacks specific to this domain with size constraints that minimize attack costs. We further discuss how these attacks can be used as an analysis tool to study and evaluate the robustness properties of financial models. Finally, we investigate the feasibility of realistic adversarial attacks in which an adversarial trader fools automated trading systems into making inaccurate predictions.

69.Towards Robust and Reproducible Active Learning Using Neural Networks ⬇️

Active learning (AL) is a promising ML paradigm that has the potential to parse through large unlabeled data and help reduce annotation cost in domains where labeling entire data can be prohibitive. Recently proposed neural network based AL methods use different heuristics to accomplish this goal. In this study, we show that recent AL methods offer a gain over random baseline under a brittle combination of experimental conditions. We demonstrate that such marginal gains vanish when experimental factors are changed, leading to reproducibility issues and suggesting that AL methods lack robustness. We also observe that with a properly tuned model, which employs recently proposed regularization techniques, the performance significantly improves for all AL methods including the random sampling baseline, and performance differences among the AL methods become negligible. Based on these observations, we suggest a set of experiments that are critical to assess the true effectiveness of an AL method. To facilitate these experiments we also present an open source toolkit. We believe our findings and recommendations will help advance reproducible research in robust AL using neural networks.

70.Self-Supervised Poisson-Gaussian Denoising ⬇️

We extend the blindspot model for self-supervised denoising to handle Poisson-Gaussian noise and introduce an improved training scheme that avoids hyperparameters and adapts the denoiser to the test data. Self-supervised models for denoising learn to denoise from only noisy data and do not require corresponding clean images, which are difficult or impossible to acquire in some application areas of interest such as low-light microscopy. We introduce a new training strategy to handle Poisson-Gaussian noise which is the standard noise model for microscope images. Our new strategy eliminates hyperparameters from the loss function, which is important in a self-supervised regime where no ground truth data is available to guide hyperparameter tuning. We show how our denoiser can be adapted to the test data to improve performance. Our evaluation on a microscope image denoising benchmark validates our approach.