1.View Generalization for Single Image Textured 3D Models ⬇️
Humans can easily infer the underlying 3D geometry and texture of an object only from a single 2D image. Current computer vision methods can do this, too, but suffer from view generalization problems - the models inferred tend to make poor predictions of appearance in novel views. As for generalization problems in machine learning, the difficulty is balancing single-view accuracy (cf. training error; bias) with novel view accuracy (cf. test error; variance). We describe a class of models whose geometric rigidity is easily controlled to manage this tradeoff. We describe a cycle consistency loss that improves view generalization (roughly, a model from a generated view should predict the original view well). View generalization of textures requires that models share texture information, so a car seen from the back still has headlights because other cars have headlights. We describe a cycle consistency loss that encourages model textures to be aligned, so as to encourage sharing. We compare our method against the state-of-the-art method and show both qualitative and quantitative improvements.
2.Step-Wise Hierarchical Alignment Network for Image-Text Matching ⬇️
Image-text matching plays a central role in bridging the semantic gap between vision and language. The key point to achieve precise visual-semantic alignment lies in capturing the fine-grained cross-modal correspondence between image and text. Most previous methods rely on single-step reasoning to discover the visual-semantic interactions, which lacks the ability of exploiting the multi-level information to locate the hierarchical fine-grained relevance. Different from them, in this work, we propose a step-wise hierarchical alignment network (SHAN) that decomposes image-text matching into multi-step cross-modal reasoning process. Specifically, we first achieve local-to-local alignment at fragment level, following by performing global-to-local and global-to-global alignment at context level sequentially. This progressive alignment strategy supplies our model with more complementary and sufficient semantic clues to understand the hierarchical correlations between image and text. The experimental results on two benchmark datasets demonstrate the superiority of our proposed method.
3.Efficient Deep Learning Architectures for Fast Identification of Bacterial Strains in Resource-Constrained Devices ⬇️
This work presents twelve fine-tuned deep learning architectures to solve the bacterial classification problem over the Digital Image of Bacterial Species Dataset. The base architectures were mainly published as mobile or efficient solutions to the ImageNet challenge, and all experiments presented in this work consisted of making several modifications to the original designs, in order to make them able to solve the bacterial classification problem by using fine-tuning and transfer learning techniques. This work also proposes a novel data augmentation technique for this dataset, which is based on the idea of artificial zooming, strongly increasing the performance of every tested architecture, even doubling it in some cases. In order to get robust and complete evaluations, all experiments were performed with 10-fold cross-validation and evaluated with five different metrics: top-1 and top-5 accuracy, precision, recall, and F1 score. This paper presents a complete comparison of the twelve different architectures, cross-validated with the original and the augmented version of the dataset, the results are also compared with several literature methods. Overall, eight of the eleven architectures surpassed the 0.95 scores in top-1 accuracy with our data augmentation method, being 0.9738 the highest top-1 accuracy. The impact of the data augmentation technique is reported with relative improvement scores.
4.Shallow Optical Flow Three-Stream CNN for Macro- and Micro-Expression Spotting from Long Videos ⬇️
Facial expressions vary from the visible to the subtle. In recent years, the analysis of micro-expressions
$-$ a natural occurrence resulting from the suppression of one's true emotions, has drawn the attention of researchers with a broad range of potential applications. However, spotting microexpressions in long videos becomes increasingly challenging when intertwined with normal or macro-expressions. In this paper, we propose a shallow optical flow three-stream CNN (SOFTNet) model to predict a score that captures the likelihood of a frame being in an expression interval. By fashioning the spotting task as a regression problem, we introduce pseudo-labeling to facilitate the learning process. We demonstrate the efficacy and efficiency of the proposed approach on the recent MEGC 2020 benchmark, where state-of-the-art performance is achieved on CAS(ME)$^{2}$ with equally promising results on SAMM Long Videos.
5.Pedestrian Attribute Recognition in Video Surveillance Scenarios Based on View-attribute Attention Localization ⬇️
Pedestrian attribute recognition in surveillance scenarios is still a challenging task due to inaccurate localization of specific attributes. In this paper, we propose a novel view-attribute localization method based on attention (VALA), which relies on the strong relevance between attributes and views to capture specific view-attributes and to localize attribute-corresponding areas by attention mechanism. A specific view-attribute is composed by the extracted attribute feature and four view scores which are predicted by view predictor as the confidences for attribute from different views. View-attribute is then delivered back to shallow network layers for supervising deep feature extraction. To explore the location of a view-attribute, regional attention is introduced to aggregate spatial information of the input attribute feature in height and width direction for constraining the image into a narrow range. Moreover, the inter-channel dependency of view-feature is embedded in the above two spatial directions. An attention attribute-specific region is gained after fining the narrow range by balancing the ratio of channel dependencies between height and width branches. The final view-attribute recognition outcome is obtained by combining the output of regional attention with the view scores from view predictor. Experiments on three wide datasets (RAP, RAPv2, PETA, and PA-100K) demonstrate the effectiveness of our approach compared with state-of-the-art methods.
6.Neural Network Modeling of Probabilities for Coding the Octree Representation of Point Clouds ⬇️
This paper describes a novel lossless point cloud compression algorithm that uses a neural network for estimating the coding probabilities for the occupancy status of voxels, depending on wide three dimensional contexts around the voxel to be encoded. The point cloud is represented as an octree, with each resolution layer being sequentially encoded and decoded using arithmetic coding, starting from the lowest resolution, until the final resolution is reached. The occupancy probability of each voxel of the splitting pattern at each node of the octree is modeled by a neural network, having at its input the already encoded occupancy status of several octree nodes (belonging to the past and current resolutions), corresponding to a 3D context surrounding the node to be encoded. The algorithm has a fast and a slow version, the fast version selecting differently several voxels of the context, which allows an increased parallelization by sending larger batches of templates to be estimated by the neural network, at both encoder and decoder. The proposed algorithms yield state-of-the-art results on benchmark datasets. The implementation will be made available at this https URL
7.K-shot NAS: Learnable Weight-Sharing for NAS with K-shot Supernets ⬇️
In one-shot weight sharing for NAS, the weights of each operation (at each layer) are supposed to be identical for all architectures (paths) in the supernet. However, this rules out the possibility of adjusting operation weights to cater for different paths, which limits the reliability of the evaluation results. In this paper, instead of counting on a single supernet, we introduce
$K$ -shot supernets and take their weights for each operation as a dictionary. The operation weight for each path is represented as a convex combination of items in a dictionary with a simplex code. This enables a matrix approximation of the stand-alone weight matrix with a higher rank ($K>1$ ). A \textit{simplex-net} is introduced to produce architecture-customized code for each path. As a result, all paths can adaptively learn how to share weights in the$K$ -shot supernets and acquire corresponding weights for better evaluation.$K$ -shot supernets and simplex-net can be iteratively trained, and we further extend the search to the channel dimension. Extensive experiments on benchmark datasets validate that K-shot NAS significantly improves the evaluation accuracy of paths and thus brings in impressive performance improvements.
8.Learning Compositional Shape Priors for Few-Shot 3D Reconstruction ⬇️
The impressive performance of deep convolutional neural networks in single-view 3D reconstruction suggests that these models perform non-trivial reasoning about the 3D structure of the output space. Recent work has challenged this belief, showing that, on standard benchmarks, complex encoder-decoder architectures perform similarly to nearest-neighbor baselines or simple linear decoder models that exploit large amounts of per-category data. However, building large collections of 3D shapes for supervised training is a laborious process; a more realistic and less constraining task is inferring 3D shapes for categories with few available training examples, calling for a model that can successfully generalize to novel object classes. In this work we experimentally demonstrate that naive baselines fail in this few-shot learning setting, in which the network must learn informative shape priors for inference of new categories. We propose three ways to learn a class-specific global shape prior, directly from data. Using these techniques, we are able to capture multi-scale information about the 3D shape, and account for intra-class variability by virtue of an implicit compositional structure. Experiments on the popular ShapeNet dataset show that our method outperforms a zero-shot baseline by over 40%, and the current state-of-the-art by over 10%, in terms of relative performance, in the few-shot setting.12
9.An Image Forensic Technique Based on JPEG Ghosts ⬇️
The unprecedented growth in the easy availability of photo-editing tools has endangered the power of digital this http URL image was supposed to be worth more than a thousand words,but now this can be said only if it can be authenticated orthe integrity of the image can be proved to be intact. In thispaper, we propose a digital image forensic technique for JPEG images. It can detect any forgery in the image if the forged portion called a ghost image is having a compression quality different from that of the cover image. It is based on resaving the JPEG image at different JPEG qualities, and the detection of the forged portion is maximum when it is saved at the same JPEG quality as the cover image. Also, we can precisely predictthe JPEG quality of the cover image by analyzing the similarity using Structural Similarity Index Measure (SSIM) or the energyof the images. The first maxima in SSIM or the first minima inenergy correspond to the cover image JPEG quality. We created adataset for varying JPEG compression qualities of the ghost and the cover images and validated the scalability of the experimental results.We also, experimented with varied attack scenarios, e.g. high-quality ghost image embedded in low quality of cover image,low-quality ghost image embedded in high-quality of cover image,and ghost image and cover image both at the same quality.The proposed method is able to localize the tampered portions accurately even for forgeries as small as 10x10 sized pixel blocks.Our technique is also robust against other attack scenarios like copy-move forgery, inserting text into image, rescaling (zoom-out/zoom-in) ghost image and then pasting on cover image.
10.A Framework to Enhance Generalization of Deep Metric Learning methods using General Discriminative Feature Learning and Class Adversarial Neural Networks ⬇️
Metric learning algorithms aim to learn a distance function that brings the semantically similar data items together and keeps dissimilar ones at a distance. The traditional Mahalanobis distance learning is equivalent to find a linear projection. In contrast, Deep Metric Learning (DML) methods are proposed that automatically extract features from data and learn a non-linear transformation from input space to a semantically embedding space. Recently, many DML methods are proposed focused to enhance the discrimination power of the learned metric by providing novel sampling strategies or loss functions. This approach is very helpful when both the training and test examples are coming from the same set of categories. However, it is less effective in many applications of DML such as image retrieval and person-reidentification. Here, the DML should learn general semantic concepts from observed classes and employ them to rank or identify objects from unseen categories. Neglecting the generalization ability of the learned representation and just emphasizing to learn a more discriminative embedding on the observed classes may lead to the overfitting problem. To address this limitation, we propose a framework to enhance the generalization power of existing DML methods in a Zero-Shot Learning (ZSL) setting by general yet discriminative representation learning and employing a class adversarial neural network. To learn a more general representation, we propose to employ feature maps of intermediate layers in a deep neural network and enhance their discrimination power through an attention mechanism. Besides, a class adversarial network is utilized to enforce the deep model to seek class invariant features for the DML task. We evaluate our work on widely used machine vision datasets in a ZSL setting.
11.Scale-invariant scale-channel networks: Deep networks that generalise to previously unseen scales ⬇️
The ability to handle large scale variations is crucial for many real world visual tasks. A straightforward approach for handling scale in a deep network is to process an image at several scales simultaneously in a set of scale channels. Scale invariance can then, in principle, be achieved by using weight sharing between the scale channels together with max or average pooling over the outputs from the scale channels. The ability of such scale channel networks to generalise to scales not present in the training set over significant scale ranges has, however, not previously been explored.
In this paper, we present a systematic study of this methodology by implementing different types of scale channel networks and evaluating their ability to generalise to previously unseen scales. We develop a formalism for analysing the covariance and invariance properties of scale channel networks, and explore how different design choices, unique to scaling transformations, affect the overall performance of scale channel networks. We first show that two previously proposed scale channel network designs do not generalise well to scales not present in the training set. We explain theoretically and demonstrate experimentally why generalisation fails in these cases.
We then propose a new type of foveated scale channel architecture}, where the scale channels process increasingly larger parts of the image with decreasing resolution. This new type of scale channel network is shown to generalise extremely well, provided sufficient image resolution and the absence of boundary effects. Our proposed FovMax and FovAvg networks perform almost identically over a scale range of 8, also when training on single scale training data, and do also give improved performance when learning from datasets with large scale variations in the small sample regime.
12.Attention-based Partial Face Recognition ⬇️
Photos of faces captured in unconstrained environments, such as large crowds, still constitute challenges for current face recognition approaches as often faces are occluded by objects or people in the foreground. However, few studies have addressed the task of recognizing partial faces. In this paper, we propose a novel approach to partial face recognition capable of recognizing faces with different occluded areas. We achieve this by combining attentional pooling of a ResNet's intermediate feature maps with a separate aggregation module. We further adapt common losses to partial faces in order to ensure that the attention maps are diverse and handle occluded parts. Our thorough analysis demonstrates that we outperform all baselines under multiple benchmark protocols, including naturally and synthetically occluded partial faces. This suggests that our method successfully focuses on the relevant parts of the occluded face.
13.Small Object Detection for Near Real-Time Egocentric Perception in a Manual Assembly Scenario ⬇️
Detecting small objects in video streams of head-worn augmented reality devices in near real-time is a huge challenge: training data is typically scarce, the input video stream can be of limited quality, and small objects are notoriously hard to detect. In industrial scenarios, however, it is often possible to leverage contextual knowledge for the detection of small objects. Furthermore, CAD data of objects are typically available and can be used to generate synthetic training data. We describe a near real-time small object detection pipeline for egocentric perception in a manual assembly scenario: We generate a training data set based on CAD data and realistic backgrounds in Unity. We then train a YOLOv4 model for a two-stage detection process: First, the context is recognized, then the small object of interest is detected. We evaluate our pipeline on the augmented reality device Microsoft Hololens 2.
14.Conterfactual Generative Zero-Shot Semantic Segmentation ⬇️
zero-shot learning is an essential part of computer vision. As a classical downstream task, zero-shot semantic segmentation has been studied because of its applicant value. One of the popular zero-shot semantic segmentation methods is based on the generative model Most new proposed works added structures on the same architecture to enhance this model. However, we found that, from the view of causal inference, the result of the original model has been influenced by spurious statistical relationships. Thus the performance of the prediction shows severe bias. In this work, we consider counterfactual methods to avoid the confounder in the original model. Based on this method, we proposed a new framework for zero-shot semantic segmentation. Our model is compared with baseline models on two real-world datasets, Pascal-VOC and Pascal-Context. The experiment results show proposed models can surpass previous confounded models and can still make use of additional structures to improve the performance. We also design a simple structure based on Graph Convolutional Networks (GCN) in this work.
15.Part-aware Panoptic Segmentation ⬇️
In this work, we introduce the new scene understanding task of Part-aware Panoptic Segmentation (PPS), which aims to understand a scene at multiple levels of abstraction, and unifies the tasks of scene parsing and part parsing. For this novel task, we provide consistent annotations on two commonly used datasets: Cityscapes and Pascal VOC. Moreover, we present a single metric to evaluate PPS, called Part-aware Panoptic Quality (PartPQ). For this new task, using the metric and annotations, we set multiple baselines by merging results of existing state-of-the-art methods for panoptic segmentation and part segmentation. Finally, we conduct several experiments that evaluate the importance of the different levels of abstraction in this single task.
16.SimSwap: An Efficient Framework For High Fidelity Face Swapping ⬇️
We propose an efficient framework, called Simple Swap (SimSwap), aiming for generalized and high fidelity face swapping. In contrast to previous approaches that either lack the ability to generalize to arbitrary identity or fail to preserve attributes like facial expression and gaze direction, our framework is capable of transferring the identity of an arbitrary source face into an arbitrary target face while preserving the attributes of the target face. We overcome the above defects in the following two ways. First, we present the ID Injection Module (IIM) which transfers the identity information of the source face into the target face at feature level. By using this module, we extend the architecture of an identity-specific face swapping algorithm to a framework for arbitrary face swapping. Second, we propose the Weak Feature Matching Loss which efficiently helps our framework to preserve the facial attributes in an implicit way. Extensive experiments on wild faces demonstrate that our SimSwap is able to achieve competitive identity performance while preserving attributes better than previous state-of-the-art methods. The code is already available on github: this https URL.
17.ViT-Inception-GAN for Image Colourising ⬇️
Studies involving colourising images has been garnering researchers' keen attention over time, assisted by significant advances in various Machine Learning techniques and compute power availability. Traditionally, colourising images have been an intricate task that gave a substantial degree of freedom during the assignment of chromatic information. In our proposed method, we attempt to colourise images using Vision Transformer - Inception - Generative Adversarial Network (ViT-I-GAN), which has an Inception-v3 fusion embedding in the generator. For a stable and robust network, we have used Vision Transformer (ViT) as the discriminator. We trained the model on the Unsplash and the COCO dataset for demonstrating the improvement made by the Inception-v3 embedding. We have compared the results between ViT-GANs with and without Inception-v3 embedding.
18.Bridge the Gap Between Model-based and Model-free Human Reconstruction ⬇️
It is challenging to directly estimate the geometry of human from a single image due to the high diversity and complexity of body shapes with the various clothing styles. Most of model-based approaches are limited to predict the shape and pose of a minimally clothed body with over-smoothing surface. Although capturing the fine detailed geometries, the model-free methods are lack of the fixed mesh topology. To address these issues, we propose a novel topology-preserved human reconstruction approach by bridging the gap between model-based and model-free human reconstruction. We present an end-to-end neural network that simultaneously predicts the pixel-aligned implicit surface and the explicit mesh model built by graph convolutional neural network. Moreover, an extra graph convolutional neural network is employed to estimate the vertex offsets between the implicit surface and parametric mesh model. Finally, we suggest an efficient implicit registration method to refine the neural network output in implicit space. Experiments on DeepHuman dataset showed that our approach is effective.
19.AugNet: End-to-End Unsupervised Visual Representation Learning with Image Augmentation ⬇️
Most of the achievements in artificial intelligence so far were accomplished by supervised learning which requires numerous annotated training data and thus costs innumerable manpower for labeling. Unsupervised learning is one of the effective solutions to overcome such difficulties. In our work, we propose AugNet, a new deep learning training paradigm to learn image features from a collection of unlabeled pictures. We develop a method to construct the similarities between pictures as distance metrics in the embedding space by leveraging the inter-correlation between augmented versions of samples. Our experiments demonstrate that the method is able to represent the image in low dimensional space and performs competitively in downstream tasks such as image classification and image similarity comparison. Specifically, we achieved over 60% and 27% accuracy on the STL10 and CIFAR100 datasets with unsupervised clustering, respectively. Moreover, unlike many deep-learning-based image retrieval algorithms, our approach does not require access to external annotated datasets to train the feature extractor, but still shows comparable or even better feature representation ability and easy-to-use characteristics. In our evaluations, the method outperforms all the state-of-the-art image retrieval algorithms on some out-of-domain image datasets. The code for the model implementation is available at this https URL.
20.A deep learning approach to clustering visual arts ⬇️
Clustering artworks is difficult for several reasons. On the one hand, recognizing meaningful patterns based on domain knowledge and visual perception is extremely hard. On the other hand, applying traditional clustering and feature reduction techniques to the highly dimensional pixel space can be ineffective. To address these issues, in this paper we propose DELIUS: a DEep learning approach to cLustering vIsUal artS. The method uses a pre-trained convolutional network to extract features and then feeds these features into a deep embedded clustering model, where the task of mapping the raw input data to a latent space is jointly optimized with the task of finding a set of cluster centroids in this latent space. Quantitative and qualitative experimental results show the effectiveness of the proposed method. DELIUS can be useful for several tasks related to art analysis, in particular visual link retrieval and historical knowledge discovery in painting datasets.
21.MlTr: Multi-label Classification with Transformer ⬇️
The task of multi-label image classification is to recognize all the object labels presented in an image. Though advancing for years, small objects, similar objects and objects with high conditional probability are still the main bottlenecks of previous convolutional neural network(CNN) based models, limited by convolutional kernels' representational capacity. Recent vision transformer networks utilize the self-attention mechanism to extract the feature of pixel granularity, which expresses richer local semantic information, while is insufficient for mining global spatial dependence. In this paper, we point out the three crucial problems that CNN-based methods encounter and explore the possibility of conducting specific transformer modules to settle them. We put forward a Multi-label Transformer architecture(MlTr) constructed with windows partitioning, in-window pixel attention, cross-window attention, particularly improving the performance of multi-label image classification tasks. The proposed MlTr shows state-of-the-art results on various prevalent multi-label datasets such as MS-COCO, Pascal-VOC, and NUS-WIDE with 88.5%, 95.8%, and 65.5% respectively. The code will be available soon at this https URL
22.Calibration and Auto-Refinement for Light Field Cameras ⬇️
The ability to create an accurate three-dimensional reconstruction of a captured scene draws attention to the principles of light fields. This paper presents an approach for light field camera calibration and rectification, based on pairwise pattern-based parameters extraction. It is followed by a correspondence-based algorithm for camera parameters refinement from arbitrary scenes using the triangulation filter and nonlinear optimization. The effectiveness of our approach is validated on both real and synthetic data.
23.Learning the Precise Feature for Cluster Assignment ⬇️
Clustering is one of the fundamental tasks in computer vision and pattern recognition. Recently, deep clustering methods (algorithms based on deep learning) have attracted wide attention with their impressive performance. Most of these algorithms combine deep unsupervised representation learning and standard clustering together. However, the separation of representation learning and clustering will lead to suboptimal solutions because the two-stage strategy prevents representation learning from adapting to subsequent tasks (e.g., clustering according to specific cues). To overcome this issue, efforts have been made in the dynamic adaption of representation and cluster assignment, whereas current state-of-the-art methods suffer from heuristically constructed objectives with representation and cluster assignment alternatively optimized. To further standardize the clustering problem, we audaciously formulate the objective of clustering as finding a precise feature as the cue for cluster assignment. Based on this, we propose a general-purpose deep clustering framework which radically integrates representation learning and clustering into a single pipeline for the first time. The proposed framework exploits the powerful ability of recently developed generative models for learning intrinsic features, and imposes an entropy minimization on the distribution of the cluster assignment by a dedicated variational algorithm. Experimental results show that the performance of the proposed method is superior, or at least comparable to, the state-of-the-art methods on the handwritten digit recognition, fashion recognition, face recognition and object recognition benchmark datasets.
24.Team RUC_AIM3 Technical Report at ActivityNet 2021: Entities Object Localization ⬇️
Entities Object Localization (EOL) aims to evaluate how grounded or faithful a description is, which consists of caption generation and object grounding. Previous works tackle this problem by jointly training the two modules in a framework, which limits the complexity of each module. Therefore, in this work, we propose to divide these two modules into two stages and improve them respectively to boost the whole system performance. For the caption generation, we propose a Unified Multi-modal Pre-training Model (UMPM) to generate event descriptions with rich objects for better localization. For the object grounding, we fine-tune the state-of-the-art detection model MDETR and design a post processing method to make the grounding results more faithful. Our overall system achieves the state-of-the-art performances on both sub-tasks in Entities Object Localization challenge at Activitynet 2021, with 72.57 localization accuracy on the testing set of sub-task I and 0.2477 F1_all_per_sent on the hidden testing set of sub-task II.
25.Refining Pseudo Labels with Clustering Consensus over Generations for Unsupervised Object Re-identification ⬇️
Unsupervised object re-identification targets at learning discriminative representations for object retrieval without any annotations. Clustering-based methods conduct training with the generated pseudo labels and currently dominate this research direction. However, they still suffer from the issue of pseudo label noise. To tackle the challenge, we propose to properly estimate pseudo label similarities between consecutive training generations with clustering consensus and refine pseudo labels with temporally propagated and ensembled pseudo labels. To the best of our knowledge, this is the first attempt to leverage the spirit of temporal ensembling to improve classification with dynamically changing classes over generations. The proposed pseudo label refinery strategy is simple yet effective and can be seamlessly integrated into existing clustering-based unsupervised re-identification methods. With our proposed approach, state-of-the-art method can be further boosted with up to 8.8% mAP improvements on the challenging MSMT17 dataset.
26.Instance-Level Task Parameters: A Robust Multi-task Weighting Framework ⬇️
Recent works have shown that deep neural networks benefit from multi-task learning by learning a shared representation across several related tasks. However, performance of such systems depend on relative weighting between various losses involved during training. Prior works on loss weighting schemes assume that instances are equally easy or hard for all tasks. In order to break this assumption, we let the training process dictate the optimal weighting of tasks for every instance in the dataset. More specifically, we equip every instance in the dataset with a set of learnable parameters (instance-level task parameters) where the cardinality is equal to the number of tasks learned by the model. These parameters model the weighting of each task for an instance. They are updated by gradient descent and do not require hand-crafted rules. We conduct extensive experiments on SURREAL and CityScapes datasets, for human shape and pose estimation, depth estimation and semantic segmentation tasks. In these tasks, our approach outperforms recent dynamic loss weighting approaches, e.g. reducing surface estimation errors by 8.97% on SURREAL. When applied to datasets where one or more tasks can have noisy annotations, the proposed method learns to prioritize learning from clean labels for a given task, e.g. reducing surface estimation errors by up to 60%. We also show that we can reliably detect corrupt labels for a given task as a by-product from learned instance-level task parameters.
27.Spectral Unsupervised Domain Adaptation for Visual Recognition ⬇️
Unsupervised domain adaptation (UDA) aims to learn a well-performed model in an unlabeled target domain by leveraging labeled data from one or multiple related source domains. It remains a great challenge due to 1) the lack of annotations in the target domain and 2) the rich discrepancy between the distributions of source and target data. We propose Spectral UDA (SUDA), an efficient yet effective UDA technique that works in the spectral space and is generic across different visual recognition tasks in detection, classification and segmentation. SUDA addresses UDA challenges from two perspectives. First, it mitigates inter-domain discrepancies by a spectrum transformer (ST) that maps source and target images into spectral space and learns to enhance domain-invariant spectra while suppressing domain-variant spectra simultaneously. To this end, we design novel adversarial multi-head spectrum attention that leverages contextual information to identify domain-variant and domain-invariant spectra effectively. Second, it mitigates the lack of annotations in target domain by introducing multi-view spectral learning which aims to learn comprehensive yet confident target representations by maximizing the mutual information among multiple ST augmentations capturing different spectral views of each target sample. Extensive experiments over different visual tasks (e.g., detection, classification and segmentation) show that SUDA achieves superior accuracy and it is also complementary with state-of-the-art UDA methods with consistent performance boosts but little extra computation.
28.A modular framework for object-based saccadic decisions in dynamic scenes ⬇️
Visually exploring the world around us is not a passive process. Instead, we actively explore the world and acquire visual information over time. Here, we present a new model for simulating human eye-movement behavior in dynamic real-world scenes. We model this active scene exploration as a sequential decision making process. We adapt the popular drift-diffusion model (DDM) for perceptual decision making and extend it towards multiple options, defined by objects present in the scene. For each possible choice, the model integrates evidence over time and a decision (saccadic eye movement) is triggered as soon as evidence crosses a decision threshold. Drawing this explicit connection between decision making and object-based scene perception is highly relevant in the context of active viewing, where decisions are made continuously while interacting with an external environment. We validate our model with a carefully designed ablation study and explore influences of our model parameters. A comparison on the VidCom dataset supports the plausibility of the proposed approach.
29.Gaussian Bounding Boxes and Probabilistic Intersection-over-Union for Object Detection ⬇️
Most object detection methods use bounding boxes to encode and represent the object shape and location. In this work, we explore a fuzzy representation of object regions using Gaussian distributions, which provides an implicit binary representation as (potentially rotated) ellipses. We also present a similarity measure for the Gaussian distributions based on the Hellinger Distance, which can be viewed as a Probabilistic Intersection-over-Union (ProbIoU). Our experimental results show that the proposed Gaussian representations are closer to annotated segmentation masks in publicly available datasets, and that loss functions based on ProbIoU can be successfully used to regress the parameters of the Gaussian representation. Furthermore, we present a simple mapping scheme from traditional (or rotated) bounding boxes to Gaussian representations, allowing the proposed ProbIoU-based losses to be seamlessly integrated into any object detector.
30.Predicting Next Local Appearance for Video Anomaly Detection ⬇️
We present a local anomaly detection method in videos. As opposed to most existing methods that are computationally expensive and are not very generalizable across different video scenes, we propose an adversarial framework that learns the temporal local appearance variations by predicting the appearance of a normally behaving object in the next frame of a scene by only relying on its current and past appearances. In the presence of an abnormally behaving object, the reconstruction error between the real and the predicted next appearance of that object indicates the likelihood of an anomaly. Our method is competitive with the existing state-of-the-art while being significantly faster for both training and inference and being better at generalizing to unseen video scenes.
31.A self-adapting super-resolution structures framework for automatic design of GAN ⬇️
With the development of deep learning, the single super-resolution image reconstruction network models are becoming more and more complex. Small changes in hyperparameters of the models have a greater impact on model performance. In the existing works, experts have gradually explored a set of optimal model parameters based on empirical values or performing brute-force search. In this paper, we introduce a new super-resolution image reconstruction generative adversarial network framework, and a Bayesian optimization method used to optimizing the hyperparameters of the generator and discriminator. The generator is made by self-calibrated convolution, and discriminator is made by convolution lays. We have defined the hyperparameters such as the number of network layers and the number of neurons. Our method adopts Bayesian optimization as a optimization policy of GAN in our model. Not only can find the optimal hyperparameter solution automatically, but also can construct a super-resolution image reconstruction network, reducing the manual workload. Experiments show that Bayesian optimization can search the optimal solution earlier than the other two optimization algorithms.
32.Overcoming Difficulty in Obtaining Dark-skinned Subjects for Remote-PPG by Synthetic Augmentation ⬇️
Camera-based remote photoplethysmography (rPPG) provides a non-contact way to measure physiological signals (e.g., heart rate) using facial videos. Recent deep learning architectures have improved the accuracy of such physiological measurement significantly, yet they are restricted by the diversity of the annotated videos. The existing datasets MMSE-HR, AFRL, and UBFC-RPPG contain roughly 10%, 0%, and 5% of dark-skinned subjects respectively. The unbalanced training sets result in a poor generalization capability to unseen subjects and lead to unwanted bias toward different demographic groups. In Western academia, it is regrettably difficult in a university setting to collect data on these dark-skinned subjects. Here we show a first attempt to overcome the lack of dark-skinned subjects by synthetic augmentation. A joint optimization framework is utilized to translate real videos from light-skinned subjects to dark skin tones while retaining their pulsatile signals. In the experiment, our method exhibits around 31% reduction in mean absolute error for the dark-skinned group and 46% improvement on bias mitigation for all the groups, as compared with the previous work trained with just real samples.
33.Scaling Vision with Sparse Mixture of Experts ⬇️
Sparsely-gated Mixture of Experts networks (MoEs) have demonstrated excellent scalability in Natural Language Processing. In Computer Vision, however, almost all performant networks are "dense", that is, every input is processed by every parameter. We present a Vision MoE (V-MoE), a sparse version of the Vision Transformer, that is scalable and competitive with the largest dense networks. When applied to image recognition, V-MoE matches the performance of state-of-the-art networks, while requiring as little as half of the compute at inference time. Further, we propose an extension to the routing algorithm that can prioritize subsets of each input across the entire batch, leading to adaptive per-image compute. This allows V-MoE to trade-off performance and compute smoothly at test-time. Finally, we demonstrate the potential of V-MoE to scale vision models, and train a 15B parameter model that attains 90.35% on ImageNet.
34.Recovery of Meteorites Using an Autonomous Drone and Machine Learning ⬇️
The recovery of freshly fallen meteorites from tracked and triangulated meteors is critical to determining their source asteroid families. However, locating meteorite fragments in strewn fields remains a challenge with very few meteorites being recovered from the meteors triangulated in past and ongoing meteor camera networks. We examined if locating meteorites can be automated using machine learning and an autonomous drone. Drones can be programmed to fly a grid search pattern and take systematic pictures of the ground over a large survey area. Those images can be analyzed using a machine learning classifier to identify meteorites in the field among many other features. Here, we describe a proof-of-concept meteorite classifier that deploys off-line a combination of different convolution neural networks to recognize meteorites from images taken by drones in the field. The system was implemented in a conceptual drone setup and tested in the suspected strewn field of a recent meteorite fall near Walker Lake, Nevada.
35.Writing by Memorizing: Hierarchical Retrieval-based Medical Report Generation ⬇️
Medical report generation is one of the most challenging tasks in medical image analysis. Although existing approaches have achieved promising results, they either require a predefined template database in order to retrieve sentences or ignore the hierarchical nature of medical report generation. To address these issues, we propose MedWriter that incorporates a novel hierarchical retrieval mechanism to automatically extract both report and sentence-level templates for clinically accurate report generation. MedWriter first employs the Visual-Language Retrieval~(VLR) module to retrieve the most relevant reports for the given images. To guarantee the logical coherence between sentences, the Language-Language Retrieval~(LLR) module is introduced to retrieve relevant sentences based on the previous generated description. At last, a language decoder fuses image features and features from retrieved reports and sentences to generate meaningful medical reports. We verified the effectiveness of our model by automatic evaluation and human evaluation on two datasets, i.e., Open-I and MIMIC-CXR.
36.Survey of Image Based Graph Neural Networks ⬇️
In this survey paper, we analyze image based graph neural networks and propose a three-step classification approach. We first convert the image into superpixels using the Quickshift algorithm so as to reduce 30% of the input data. The superpixels are subsequently used to generate a region adjacency graph. Finally, the graph is passed through a state-of-art graph convolutional neural network to get classification scores. We also analyze the spatial and spectral convolution filtering techniques in graph neural networks. Spectral-based models perform better than spatial-based models and classical CNN with lesser compute cost.
37.KRADA: Known-region-aware Domain Alignment for Open World Semantic Segmentation ⬇️
In semantic segmentation, we aim to train a pixel-level classifier to assign category labels to all pixels in an image, where labeled training images and unlabeled test images are from the same distribution and share the same label set. However, in an open world, the unlabeled test images probably contain unknown categories and have different distributions from the labeled images. Hence, in this paper, we consider a new, more realistic, and more challenging problem setting where the pixel-level classifier has to be trained with labeled images and unlabeled open-world images -- we name it open world semantic segmentation (OSS). In OSS, the trained classifier is expected to identify unknown-class pixels and classify known-class pixels well. To solve OSS, we first investigate which distribution that unknown-class pixels obey. Then, motivated by the goodness-of-fit test, we use statistical measurements to show how a pixel fits the distribution of an unknown class and select highly-fitted pixels to form the unknown region in each image. Eventually, we propose an end-to-end learning framework, known-region-aware domain alignment (KRADA), to distinguish unknown classes while aligning distributions of known classes in labeled and unlabeled open-world images. The effectiveness of KRADA has been verified on two synthetic tasks and one COVID-19 segmentation task.
38.PyGAD: An Intuitive Genetic Algorithm Python Library ⬇️
This paper introduces PyGAD, an open-source easy-to-use Python library for building the genetic algorithm. PyGAD supports a wide range of parameters to give the user control over everything in its life cycle. This includes, but is not limited to, population, gene value range, gene data type, parent selection, crossover, and mutation. PyGAD is designed as a general-purpose optimization library that allows the user to customize the fitness function. Its usage consists of 3 main steps: build the fitness function, create an instance of the this http URL class, and calling the pygad.GA.run() method. The library supports training deep learning models created either with PyGAD itself or with frameworks like Keras and PyTorch. Given its stable state, PyGAD is also in active development to respond to the user's requested features and enhancement received on GitHub this https URL. PyGAD comes with documentation this https URL for further details and examples.
39.Progressive-Scale Boundary Blackbox Attack via Projective Gradient Estimation ⬇️
Boundary based blackbox attack has been recognized as practical and effective, given that an attacker only needs to access the final model prediction. However, the query efficiency of it is in general high especially for high dimensional image data. In this paper, we show that such efficiency highly depends on the scale at which the attack is applied, and attacking at the optimal scale significantly improves the efficiency. In particular, we propose a theoretical framework to analyze and show three key characteristics to improve the query efficiency. We prove that there exists an optimal scale for projective gradient estimation. Our framework also explains the satisfactory performance achieved by existing boundary black-box attacks. Based on our theoretical framework, we propose Progressive-Scale enabled projective Boundary Attack (PSBA) to improve the query efficiency via progressive scaling techniques. In particular, we employ Progressive-GAN to optimize the scale of projections, which we call PSBA-PGAN. We evaluate our approach on both spatial and frequency scales. Extensive experiments on MNIST, CIFAR-10, CelebA, and ImageNet against different models including a real-world face recognition API show that PSBA-PGAN significantly outperforms existing baseline attacks in terms of query efficiency and attack success rate. We also observe relatively stable optimal scales for different models and datasets. The code is publicly available at this https URL.
40.Rethinking Architecture Design for Tackling Data Heterogeneity in Federated Learning ⬇️
Federated learning is an emerging research paradigm enabling collaborative training of machine learning models among different organizations while keeping data private at each institution. Despite recent progress, there remain fundamental challenges such as lack of convergence and potential for catastrophic forgetting in federated learning across real-world heterogeneous devices. In this paper, we demonstrate that attention-based architectures (e.g., Transformers) are fairly robust to distribution shifts and hence improve federated learning over heterogeneous data. Concretely, we conduct the first rigorous empirical investigation of different neural architectures across a range of federated algorithms, real-world benchmarks, and heterogeneous data splits. Our experiments show that simply replacing convolutional networks with Transformers can greatly reduce catastrophic forgetting of previous devices, accelerate convergence, and reach a better global model, especially when dealing with heterogeneous data. We will release our code and pretrained models at this https URL to encourage future exploration in robust architectures as an alternative to current research efforts on the optimization front.
41.Sparse and Imperceptible Adversarial Attack via a Homotopy Algorithm ⬇️
Sparse adversarial attacks can fool deep neural networks (DNNs) by only perturbing a few pixels (regularized by l_0 norm). Recent efforts combine it with another l_infty imperceptible on the perturbation magnitudes. The resultant sparse and imperceptible attacks are practically relevant, and indicate an even higher vulnerability of DNNs that we usually imagined. However, such attacks are more challenging to generate due to the optimization difficulty by coupling the l_0 regularizer and box constraints with a non-convex objective. In this paper, we address this challenge by proposing a homotopy algorithm, to jointly tackle the sparsity and the perturbation bound in one unified framework. Each iteration, the main step of our algorithm is to optimize an l_0-regularized adversarial loss, by leveraging the nonmonotone Accelerated Proximal Gradient Method (nmAPG) for nonconvex programming; it is followed by an l_0 change control step, and an optional post-attack step designed to escape bad local minima. We also extend the algorithm to handling the structural sparsity regularizer. We extensively examine the effectiveness of our proposed homotopy attack for both targeted and non-targeted attack scenarios, on CIFAR-10 and ImageNet datasets. Compared to state-of-the-art methods, our homotopy attack leads to significantly fewer perturbations, e.g., reducing 42.91% on CIFAR-10 and 75.03% on ImageNet (average case, targeted attack), at similar maximal perturbation magnitudes, when still achieving 100% attack success rates. Our codes are available at: this https URL.
42.Coordinate Independent Convolutional Networks -- Isometry and Gauge Equivariant Convolutions on Riemannian Manifolds ⬇️
Motivated by the vast success of deep convolutional networks, there is a great interest in generalizing convolutions to non-Euclidean manifolds. A major complication in comparison to flat spaces is that it is unclear in which alignment a convolution kernel should be applied on a manifold. The underlying reason for this ambiguity is that general manifolds do not come with a canonical choice of reference frames (gauge). Kernels and features therefore have to be expressed relative to arbitrary coordinates. We argue that the particular choice of coordinatization should not affect a network's inference -- it should be coordinate independent. A simultaneous demand for coordinate independence and weight sharing is shown to result in a requirement on the network to be equivariant under local gauge transformations (changes of local reference frames). The ambiguity of reference frames depends thereby on the G-structure of the manifold, such that the necessary level of gauge equivariance is prescribed by the corresponding structure group G. Coordinate independent convolutions are proven to be equivariant w.r.t. those isometries that are symmetries of the G-structure. The resulting theory is formulated in a coordinate free fashion in terms of fiber bundles. To exemplify the design of coordinate independent convolutions, we implement a convolutional network on the Möbius strip. The generality of our differential geometric formulation of convolutional networks is demonstrated by an extensive literature review which explains a large number of Euclidean CNNs, spherical CNNs and CNNs on general surfaces as specific instances of coordinate independent convolutions.
43.Within-layer Diversity Reduces Generalization Gap ⬇️
Neural networks are composed of multiple layers arranged in a hierarchical structure jointly trained with a gradient-based optimization, where the errors are back-propagated from the last layer back to the first one. At each optimization step, neurons at a given layer receive feedback from neurons belonging to higher layers of the hierarchy. In this paper, we propose to complement this traditional 'between-layer' feedback with additional 'within-layer' feedback to encourage diversity of the activations within the same layer. To this end, we measure the pairwise similarity between the outputs of the neurons and use it to model the layer's overall diversity. By penalizing similarities and promoting diversity, we encourage each neuron to learn a distinctive representation and, thus, to enrich the data representation learned within the layer and to increase the total capacity of the model. We theoretically study how the within-layer activation diversity affects the generalization performance of a neural network and prove that increasing the diversity of hidden activations reduces the estimation error. In addition to the theoretical guarantees, we present an empirical study on three datasets confirming that the proposed approach enhances the performance of state-of-the-art neural network models and decreases the generalization gap.