20201217.md

ArXiv cs.CV --Thu, 17 Dec 2020

1.Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contexts ⬇️

The rapid progress in 3D scene understanding has come with growing demand for data; however, collecting and annotating 3D scenes (e.g. point clouds) are notoriously hard. For example, the number of scenes (e.g. indoor rooms) that can be accessed and scanned might be limited; even given sufficient data, acquiring 3D labels (e.g. instance masks) requires intensive human labor. In this paper, we explore data-efficient learning for 3D point cloud. As a first step towards this direction, we propose Contrastive Scene Contexts, a 3D pre-training method that makes use of both point-level correspondences and spatial contexts in a scene. Our method achieves state-of-the-art results on a suite of benchmarks where training data or labels are scarce. Our study reveals that exhaustive labelling of 3D point clouds might be unnecessary; and remarkably, on ScanNet, even using 0.1% of point labels, we still achieve 89% (instance segmentation) and 96% (semantic segmentation) of the baseline performance that uses full annotations.

2.Point Transformer ⬇️

Self-attention networks have revolutionized natural language processing and are making impressive strides in image analysis tasks such as image classification and object detection. Inspired by this success, we investigate the application of self-attention networks to 3D point cloud processing. We design self-attention layers for point clouds and use these to construct self-attention networks for tasks such as semantic scene segmentation, object part segmentation, and object classification. Our Point Transformer design improves upon prior work across domains and tasks. For example, on the challenging S3DIS dataset for large-scale semantic scene segmentation, the Point Transformer attains an mIoU of 70.4% on Area 5, outperforming the strongest prior model by 3.3 absolute percentage points and crossing the 70% mIoU threshold for the first time.

3.Learning Continuous Image Representation with Local Implicit Image Function ⬇️

How to represent an image? While the visual world is presented in a continuous manner, machines store and see the images in a discrete way with 2D arrays of pixels. In this paper, we seek to learn a continuous representation for images. Inspired by the recent progress in 3D reconstruction with implicit function, we propose Local Implicit Image Function (LIIF), which takes an image coordinate and the 2D deep features around the coordinate as inputs, predicts the RGB value at a given coordinate as an output. Since the coordinates are continuous, LIIF can be presented in an arbitrary resolution. To generate the continuous representation for pixel-based images, we train an encoder and LIIF representation via a self-supervised task with super-resolution. The learned continuous representation can be presented in arbitrary resolution even extrapolate to $\times 30$ higher resolution, where the training tasks are not provided. We further show that LIIF representation builds a bridge between discrete and continuous representation in 2D, it naturally supports the learning tasks with size-varied image ground-truths and significantly outperforms the method with resizing the ground-truths. Our project page with code is at this https URL .

4.DECOR-GAN: 3D Shape Detailization by Conditional Refinement ⬇️

We introduce a deep generative network for 3D shape detailization, akin to stylization with the style being geometric details. We address the challenge of creating large varieties of high-resolution and detailed 3D geometry from a small set of exemplars by treating the problem as that of geometric detail transfer. Given a low-resolution coarse voxel shape, our network refines it, via voxel upsampling, into a higher-resolution shape enriched with geometric details. The output shape preserves the overall structure (or content) of the input, while its detail generation is conditioned on an input "style code" corresponding to a detailed exemplar. Our 3D detailization via conditional refinement is realized by a generative adversarial network, coined DECOR-GAN. The network utilizes a 3D CNN generator for upsampling coarse voxels and a 3D PatchGAN discriminator to enforce local patches of the generated model to be similar to those in the training detailed shapes. During testing, a style code is fed into the generator to condition the refinement. We demonstrate that our method can refine a coarse shape into a variety of detailed shapes with different styles. The generated results are evaluated in terms of content preservation, plausibility, and diversity. Comprehensive ablation studies are conducted to validate our network designs.

5.Joint Generative and Contrastive Learning for Unsupervised Person Re-identification ⬇️

Annotating identity labels in large-scale datasets is a labour-intensive work, which strongly limits the scalability of person re-identification (ReID) in the real world. Unsupervised ReID addresses this issue by learning representations directly from unlabeled images. Recent self-supervised contrastive learning provides an effective approach for unsupervised representation learning. In this paper, we incorporate a Generative Adversarial Network (GAN) and contrastive learning into one joint training framework. While the GAN provides online data augmentation for contrastive learning, the contrastive module learns view-invariant features for generation. In this context, we propose a mesh-based novel view generator. Specifically, mesh projections serve as references towards generating novel views of a person. In addition, we propose a view-invariant loss to facilitate contrastive learning between original and generated views. Deviating from previous GAN-based unsupervised ReID methods involving domain adaptation, we do not rely on a labeled source dataset, which makes our method more flexible. Extensive experimental results show that our method significantly outperforms state-of-the-art methods under both, fully unsupervised and unsupervised domain adaptive settings on several large scale ReID datsets.

6.Towards Recognizing New Semantic Concepts in New Visual Domains ⬇️

Deep learning models heavily rely on large scale annotated datasets for training. Unfortunately, datasets cannot capture the infinite variability of the real world, thus neural networks are inherently limited by the restricted visual and semantic information contained in their training set. In this thesis, we argue that it is crucial to design deep architectures that can operate in previously unseen visual domains and recognize novel semantic concepts. In the first part of the thesis, we describe different solutions to enable deep models to generalize to new visual domains, by transferring knowledge from a labeled source domain(s) to a domain (target) where no labeled data are available. We will show how variants of batch-normalization (BN) can be applied to different scenarios, from domain adaptation when source and target are mixtures of multiple latent domains, to domain generalization, continuous domain adaptation, and predictive domain adaptation, where information about the target domain is available only in the form of metadata. In the second part of the thesis, we show how to extend the knowledge of a pretrained deep model to new semantic concepts, without access to the original training set. We address the scenarios of sequential multi-task learning, using transformed task-specific binary masks, open-world recognition, with end-to-end training and enforced clustering, and incremental class learning in semantic segmentation, where we highlight and address the problem of the semantic shift of the background class. In the final part, we tackle a more challenging problem: given images of multiple domains and semantic categories (with their attributes), how to build a model that recognizes images of unseen concepts in unseen domains? We also propose an approach based on domain and semantic mixing of inputs and features, which is a first, promising step towards solving this problem.

7.Improved StyleGAN Embedding: Where are the Good Latents? ⬇️

StyleGAN is able to produce photorealistic images almost indistinguishable from real ones. Embedding images into the StyleGAN latent space is not a trivial task due to the reconstruction quality and editing quality trade-off. In this paper, we first introduce a new normalized space to analyze the diversity and the quality of the reconstructed latent codes. This space can help answer the question of where good latent codes are located in latent space. Second, we propose a framework to analyze the quality of different embedding algorithms. Third, we propose an improved embedding algorithm based on our analysis. We compare our results with the current state-of-the-art methods and achieve a better trade-off between reconstruction quality and editing quality.

8.CompositeTasking: Understanding Images by Spatial Composition of Tasks ⬇️

We define the concept of CompositeTasking as the fusion of multiple, spatially distributed tasks, for various aspects of image understanding. Learning to perform spatially distributed tasks is motivated by the frequent availability of only sparse labels across tasks, and the desire for a compact multi-tasking network. To facilitate CompositeTasking, we introduce a novel task conditioning model -- a single encoder-decoder network that performs multiple, spatially varying tasks at once. The proposed network takes a pair of an image and a set of pixel-wise dense tasks as inputs, and makes the task related predictions for each pixel, which includes the decision of applying which task where. As to the latter, we learn the composition of tasks that needs to be performed according to some CompositeTasking rules. It not only offers us a compact network for multi-tasking, but also allows for task-editing. The strength of the proposed method is demonstrated by only having to supply sparse supervision per task. The obtained results are on par with our baselines that use dense supervision and a multi-headed multi-tasking design. The source code will be made publicly available at this http URL .

9.AdjointBackMap: Reconstructing Effective Decision Hypersurfaces from CNN Layers Using Adjoint Operators ⬇️

There are several effective methods in explaining the inner workings of convolutional neural networks (CNNs). However, in general, finding the inverse of the function performed by CNNs as a whole is an ill-posed problem. In this paper, we propose a method based on adjoint operators to reconstruct, given an arbitrary unit in the CNN (except for the first convolutional layer), its effective hypersurface in the input space that replicates that unit's decision surface conditioned on a particular input image. Our results show that the hypersurface reconstructed this way, when multiplied by the original input image, would give nearly exact output value of that unit. We find that the CNN unit's decision surface is largely conditioned on the input, and this may explain why adversarial inputs can effectively deceive CNNs.

10.I3DOL: Incremental 3D Object Learning without Catastrophic Forgetting ⬇️

3D object classification has attracted appealing attentions in academic researches and industrial applications. However, most existing methods need to access the training data of past 3D object classes when facing the common real-world scenario: new classes of 3D objects arrive in a sequence. Moreover, the performance of advanced approaches degrades dramatically for past learned classes (i.e., catastrophic forgetting), due to the irregular and redundant geometric structures of 3D point cloud data. To address these challenges, we propose a new Incremental 3D Object Learning (i.e., I3DOL) model, which is the first exploration to learn new classes of 3D object continually. Specifically, an adaptive-geometric centroid module is designed to construct discriminative local geometric structures, which can better characterize the irregular point cloud representation for 3D object. Afterwards, to prevent the catastrophic forgetting brought by redundant geometric information, a geometric-aware attention mechanism is developed to quantify the contributions of local geometric structures, and explore unique 3D geometric characteristics with high contributions for classes incremental learning. Meanwhile, a score fairness compensation strategy is proposed to further alleviate the catastrophic forgetting caused by unbalanced data between past and new classes of 3D object, by compensating biased prediction for new classes in the validation phase. Experiments on 3D representative datasets validate the superiority of our I3DOL framework.

11.Sketch Generation with Drawing Process Guided by Vector Flow and Grayscale ⬇️

We propose a novel image-to-pencil translation method that could not only generate high-quality pencil sketches but also offer the drawing process. Existing pencil sketch algorithms are based on texture rendering rather than the direct imitation of strokes, making them unable to show the drawing process but only a final result. To address this challenge, we first establish a pencil stroke imitation mechanism. Next, we develop a framework with three branches to guide stroke drawing: the first branch guides the direction of the strokes, the second branch determines the shade of the strokes, and the third branch enhances the details further. Under this framework's guidance, we can produce a pencil sketch by drawing one stroke every time. Our method is fully interpretable. Comparison with existing pencil drawing algorithms shows that our method is superior to others in terms of texture quality, style, and user evaluation.

12.C2F-FWN: Coarse-to-Fine Flow Warping Network for Spatial-Temporal Consistent Motion Transfer ⬇️

Human video motion transfer (HVMT) aims to synthesize videos that one person imitates other persons' actions. Although existing GAN-based HVMT methods have achieved great success, they either fail to preserve appearance details due to the loss of spatial consistency between synthesized and exemplary images, or generate incoherent video results due to the lack of temporal consistency among video frames. In this paper, we propose Coarse-to-Fine Flow Warping Network (C2F-FWN) for spatial-temporal consistent HVMT. Particularly, C2F-FWN utilizes coarse-to-fine flow warping and Layout-Constrained Deformable Convolution (LC-DConv) to improve spatial consistency, and employs Flow Temporal Consistency (FTC) Loss to enhance temporal consistency. In addition, provided with multi-source appearance inputs, C2F-FWN can support appearance attribute editing with great flexibility and efficiency. Besides public datasets, we also collected a large-scale HVMT dataset named SoloDance for evaluation. Extensive experiments conducted on our SoloDance dataset and the iPER dataset show that our approach outperforms state-of-art HVMT methods in terms of both spatial and temporal consistency. Source code and the SoloDance dataset are available at this https URL.

13.SimuGAN: Unsupervised forward modeling and optimal design of a LIDAR Camera ⬇️

Energy-saving LIDAR camera for short distances estimates an object's distance using temporally intensity-coded laser light pulses and calculates the maximum correlation with the back-scattered pulse.
Though on low power, the backs-scattered pulse is noisy and unstable, which leads to inaccurate and unreliable depth estimation.
To address this problem, we use GANs (Generative Adversarial Networks), which are two neural networks that can learn complicated class distributions through an adversarial process. We learn the LIDAR camera's hidden properties and behavior, creating a novel, fully unsupervised forward model that simulates the camera. Then, we use the model's differentiability to explore the camera parameter space and optimize those parameters in terms of depth, accuracy, and stability. To achieve this goal, we also propose a new custom loss function designated to the back-scattered code distribution's weaknesses and its circular behavior. The results are demonstrated on both synthetic and real data.

14.Deep Reinforcement Learning of Graph Matching ⬇️

Graph matching under node and pairwise constraints has been a building block in areas from combinatorial optimization, machine learning to computer vision, for effective structural representation and association. We present a reinforcement learning solver that seeks the node correspondence between two graphs, whereby the node embedding model on the association graph is learned to sequentially find the node-to-node matching. Our method differs from the previous deep graph matching model in the sense that they are focused on the front-end feature and affinity function learning while our method aims to learn the backend decision making given the affinity objective function whatever obtained by learning or not. Such an objective function maximization setting naturally fits with the reinforcement learning mechanism, of which the learning procedure is label-free. Besides, the model is not restricted to a fixed number of nodes for matching. These features make it more suitable for practical usage. Extensive experimental results on both synthetic datasets, natural images, and QAPLIB showcase the superior performance regarding both matching accuracy and efficiency. To our best knowledge, this is the first deep reinforcement learning solver for graph matching.

15.SAfE: Self-Attention Based Unsupervised Road Safety Classification in Hazardous Environments ⬇️

We present a novel approach SAfE that can identify parts of an outdoor scene that are safe for driving, based on attention models. Our formulation is designed for hazardous weather conditions that can impair the visibility of human drivers as well as autonomous vehicles, increasing the risk of accidents. Our approach is unsupervised and uses domain adaptation, with entropy minimization and attention transfer discriminators, to leverage the large amounts of labeled data corresponding to clear weather conditions. Our attention transfer discriminator uses attention maps from the clear weather image to help the network learn relevant regions to attend to, on the images from the hazardous weather dataset. We conduct experiments on CityScapes simulated datasets depicting various weather conditions such as rain, fog and snow under different intensities, and additionally on Berkeley Deep Drive. Our result show that using attention models improves the standard unsupervised domain adaptation performance by 29.29%. Furthermore, we also compare with unsupervised domain adaptation methods and show an improvement of at least 12.02% (mIoU) over the state-of-the-art.

16.Copyspace: Where to Write on Images? ⬇️

The placement of text over an image is an important part of producing high-quality visual designs. Automating this work by determining appropriate position, orientation, and style for textual elements requires understanding the contents of the background image. We refer to the search for aesthetic parameters of text rendered over images as "copyspace detection", noting that this task is distinct from foreground-background separation. We have developed solutions using one and two stage object detection methodologies trained on an expertly labeled data. This workshop will examine such algorithms for copyspace detection and demonstrate their application in generative design models and pipelines such as Einstein Designer.

17.FuseVis: Interpreting neural networks for image fusion using per-pixel saliency visualization ⬇️

Image fusion helps in merging two or more images to construct a more informative single fused image. Recently, unsupervised learning based convolutional neural networks (CNN) have been utilized for different types of image fusion tasks such as medical image fusion, infrared-visible image fusion for autonomous driving as well as multi-focus and multi-exposure image fusion for satellite imagery. However, it is challenging to analyze the reliability of these CNNs for the image fusion tasks since no groundtruth is available. This led to the use of a wide variety of model architectures and optimization functions yielding quite different fusion results. Additionally, due to the highly opaque nature of such neural networks, it is difficult to explain the internal mechanics behind its fusion results. To overcome these challenges, we present a novel real-time visualization tool, named FuseVis, with which the end-user can compute per-pixel saliency maps that examine the influence of the input image pixels on each pixel of the fused image. We trained several image fusion based CNNs on medical image pairs and then using our FuseVis tool, we performed case studies on a specific clinical application by interpreting the saliency maps from each of the fusion methods. We specifically visualized the relative influence of each input image on the predictions of the fused image and showed that some of the evaluated image fusion methods are better suited for the specific clinical application. To the best of our knowledge, currently, there is no approach for visual analysis of neural networks for image fusion. Therefore, this work opens up a new research direction to improve the interpretability of deep fusion networks. The FuseVis tool can also be adapted in other deep neural network based image processing applications to make them interpretable.

18.Unsupervised Image Segmentation using Mutual Mean-Teaching ⬇️

Unsupervised image segmentation aims at assigning the pixels with similar feature into a same cluster without annotation, which is an important task in computer vision. Due to lack of prior knowledge, most of existing model usually need to be trained several times to obtain suitable results. To address this problem, we propose an unsupervised image segmentation model based on the Mutual Mean-Teaching (MMT) framework to produce more stable results. In addition, since the labels of pixels from two model are not matched, a label alignment algorithm based on the Hungarian algorithm is proposed to match the cluster labels. Experimental results demonstrate that the proposed model is able to segment various types of images and achieves better performance than the existing methods.

19.Self-Supervised Person Detection in 2D Range Data using a Calibrated Camera ⬇️

Deep learning is the essential building block of state-of-the-art person detectors in 2D range data. However, only a few annotated datasets are available for training and testing these deep networks, potentially limiting their performance when deployed in new environments or with different LiDAR models. We propose a method, which uses bounding boxes from an image-based detector (e.g. Faster R-CNN) on a calibrated camera to automatically generate training labels (called pseudo-labels) for 2D LiDAR-based person detectors. Through experiments on the JackRabbot dataset with two detector models, DROW3 and DR-SPAAM, we show that self-supervised detectors, trained or fine-tuned with pseudo-labels, outperform detectors trained using manual annotations from a different dataset. Combined with robust training techniques, the self-supervised detectors reach a performance close to the ones trained using manual annotations. Our method is an effective way to improve person detectors during deployment without any additional labeling effort, and we release our source code to support relevant robotic applications.

20.Temporal Graph Modeling for Skeleton-based Action Recognition ⬇️

Graph Convolutional Networks (GCNs), which model skeleton data as graphs, have obtained remarkable performance for skeleton-based action recognition. Particularly, the temporal dynamic of skeleton sequence conveys significant information in the recognition task. For temporal dynamic modeling, GCN-based methods only stack multi-layer 1D local convolutions to extract temporal relations between adjacent time steps. With the repeat of a lot of local convolutions, the key temporal information with non-adjacent temporal distance may be ignored due to the information dilution. Therefore, these methods still remain unclear how to fully explore temporal dynamic of skeleton sequence. In this paper, we propose a Temporal Enhanced Graph Convolutional Network (TE-GCN) to tackle this limitation. The proposed TE-GCN constructs temporal relation graph to capture complex temporal dynamic. Specifically, the constructed temporal relation graph explicitly builds connections between semantically related temporal features to model temporal relations between both adjacent and non-adjacent time steps. Meanwhile, to further explore the sufficient temporal dynamic, multi-head mechanism is designed to investigate multi-kinds of temporal relations. Extensive experiments are performed on two widely used large-scale datasets, NTU-60 RGB+D and NTU-120 RGB+D. And experimental results show that the proposed model achieves the state-of-the-art performance by making contribution to temporal modeling for action recognition.

21.Latent Space Conditioning on Generative Adversarial Networks ⬇️

Generative adversarial networks are the state of the art approach towards learned synthetic image generation. Although early successes were mostly unsupervised, bit by bit, this trend has been superseded by approaches based on labelled data. These supervised methods allow a much finer-grained control of the output image, offering more flexibility and stability. Nevertheless, the main drawback of such models is the necessity of annotated data. In this work, we introduce an novel framework that benefits from two popular learning techniques, adversarial training and representation learning, and takes a step towards unsupervised conditional GANs. In particular, our approach exploits the structure of a latent space (learned by the representation learning) and employs it to condition the generative model. In this way, we break the traditional dependency between condition and label, substituting the latter by unsupervised features coming from the latent space. Finally, we show that this new technique is able to produce samples on demand keeping the quality of its supervised counterpart.

22.Analysing the Direction of Emotional Influence in Nonverbal Dyadic Communication: A Facial-Expression Study ⬇️

Identifying the direction of emotional influence in a dyadic dialogue is of increasing interest in the psychological sciences with applications in psychotherapy, analysis of political interactions, or interpersonal conflict behavior. Facial expressions are widely described as being automatic and thus hard to overtly influence. As such, they are a perfect measure for a better understanding of unintentional behavior cues about social-emotional cognitive processes. With this view, this study is concerned with the analysis of the direction of emotional influence in dyadic dialogue based on facial expressions only. We exploit computer vision capabilities along with causal inference theory for quantitative verification of hypotheses on the direction of emotional influence, i.e., causal effect relationships, in dyadic dialogues. We address two main issues. First, in a dyadic dialogue, emotional influence occurs over transient time intervals and with intensity and direction that are variant over time. To this end, we propose a relevant interval selection approach that we use prior to causal inference to identify those transient intervals where causal inference should be applied. Second, we propose to use fine-grained facial expressions that are present when strong distinct facial emotions are not visible. To specify the direction of influence, we apply the concept of Granger causality to the time series of facial expressions over selected relevant intervals. We tested our approach on newly, experimentally obtained data. Based on the quantitative verification of hypotheses on the direction of emotional influence, we were able to show that the proposed approach is most promising to reveal the causal effect pattern in various instructed interaction conditions.

23.Revisiting 3D Context Modeling with Supervised Pre-training for Universal Lesion Detection in CT Slices ⬇️

Universal lesion detection from computed tomography (CT) slices is important for comprehensive disease screening. Since each lesion can locate in multiple adjacent slices, 3D context modeling is of great significance for developing automated lesion detection algorithms. In this work, we propose a Modified Pseudo-3D Feature Pyramid Network (MP3D FPN) that leverages depthwise separable convolutional filters and a group transform module (GTM) to efficiently extract 3D context enhanced 2D features for universal lesion detection in CT slices. To facilitate faster convergence, a novel 3D network pre-training method is derived using solely large-scale 2D object detection dataset in the natural image domain. We demonstrate that with the novel pre-training method, the proposed MP3D FPN achieves state-of-the-art detection performance on the DeepLesion dataset (3.48% absolute improvement in the sensitivity of FPs@0.5), significantly surpassing the baseline method by up to 6.06% (in MAP@0.5) which adopts 2D convolution for 3D context modeling. Moreover, the proposed 3D pre-trained weights can potentially be used to boost the performance of other 3D medical image analysis tasks.

24.Difficulty in estimating visual information from randomly sampled images ⬇️

In this paper, we evaluate dimensionality reduction methods in terms of difficulty in estimating visual information on original images from dimensionally reduced ones. Recently, dimensionality reduction has been receiving attention as the process of not only reducing the number of random variables, but also protecting visual information for privacy-preserving machine learning. For such a reason, difficulty in estimating visual information is discussed. In particular, the random sampling method that was proposed for privacy-preserving machine learning, is compared with typical dimensionality reduction methods. In an image classification experiment, the random sampling method is demonstrated not only to have high difficulty, but also to be comparable to other dimensionality reduction methods, while maintaining the property of spatial information invariant.

25.Exploiting Sample Uncertainty for Domain Adaptive Person Re-Identification ⬇️

Many unsupervised domain adaptive (UDA) person re-identification (ReID) approaches combine clustering-based pseudo-label prediction with feature fine-tuning. However, because of domain gap, the pseudo-labels are not always reliable and there are noisy/incorrect labels. This would mislead the feature representation learning and deteriorate the performance. In this paper, we propose to estimate and exploit the credibility of the assigned pseudo-label of each sample to alleviate the influence of noisy labels, by suppressing the contribution of noisy samples. We build our baseline framework using the mean teacher method together with an additional contrastive loss. We have observed that a sample with a wrong pseudo-label through clustering in general has a weaker consistency between the output of the mean teacher model and the student model. Based on this finding, we propose to exploit the uncertainty (measured by consistency levels) to evaluate the reliability of the pseudo-label of a sample and incorporate the uncertainty to re-weight its contribution within various ReID losses, including the identity (ID) classification loss per sample, the triplet loss, and the contrastive loss. Our uncertainty-guided optimization brings significant improvement and achieves the state-of-the-art performance on benchmark datasets.

26.Event-based Motion Segmentation with Spatio-Temporal Graph Cuts ⬇️

Identifying independently moving objects is an essential task for dynamic scene understanding. However, traditional cameras used in dynamic scenes may suffer from motion blur or exposure artifacts due to their sampling principle. By contrast, event-based cameras are novel bio-inspired sensors that offer advantages to overcome such limitations. They report pixel-wise intensity changes asynchronously, which enables them to acquire visual information at exactly the same rate as the scene dynamics. We have developed a method to identify independently moving objects acquired with an event-based camera, i.e., to solve the event-based motion segmentation problem. This paper describes how to formulate the problem as a weakly-constrained multi-model fitting one via energy minimization, and how to jointly solve its two subproblems -- event-cluster assignment (labeling) and motion model fitting -- in an iterative manner, by exploiting the spatio-temporal structure of input events in the form of a space-time graph. Experiments on available datasets demonstrate the versatility of the method in scenes with different motion patterns and number of moving objects. The evaluation shows that the method performs on par or better than the state of the art without having to predetermine the number of expected moving objects.

27.Deep Learning to Segment Pelvic Bones: Large-scale CT Datasets and Baseline Models ⬇️

Purpose: Pelvic bone segmentation in CT has always been an essential step in clinical diagnosis and surgery planning of pelvic bone diseases. Existing methods for pelvic bone segmentation are either hand-crafted or semi-automatic and achieve limited accuracy when dealing with image appearance variations due to the multi-site domain shift, the presence of contrasted vessels, coprolith and chyme, bone fractures, low dose, metal artifacts, etc. Due to the lack of a large-scale pelvic CT dataset with annotations, deep learning methods are not fully explored. Methods: In this paper, we aim to bridge the data gap by curating a large pelvic CT dataset pooled from multiple sources and different manufacturers, including 1, 184 CT volumes and over 320, 000 slices with different resolutions and a variety of the above-mentioned appearance variations. Then we propose for the first time, to the best of our knowledge, to learn a deep multi-class network for segmenting lumbar spine, sacrum, left hip, and right hip, from multiple-domain images simultaneously to obtain more effective and robust feature representations. Finally, we introduce a post-processing tool based on the signed distance function (SDF) to eliminate false predictions while retaining correctly predicted bone fragments. Results: Extensive experiments on our dataset demonstrate the effectiveness of our automatic method, achieving an average Dice of 0.987 for a metal-free volume. SDF post-processor yields a decrease of 10.5% in hausdorff distance by maintaining important bone fragments in post-processing phase. Conclusion: We believe this large-scale dataset will promote the development of the whole community and plan to open source the images, annotations, codes, and trained baseline models at this URL1.

28.SID-NISM: A Self-supervised Low-light Image Enhancement Framework ⬇️

When capturing images in low-light conditions, the images often suffer from low visibility, which not only degrades the visual aesthetics of images, but also significantly degenerates the performance of many computer vision algorithms. In this paper, we propose a self-supervised low-light image enhancement framework (SID-NISM), which consists of two components, a Self-supervised Image Decomposition Network (SID-Net) and a Nonlinear Illumination Saturation Mapping function (NISM). As a self-supervised network, SID-Net could decompose the given low-light image into its reflectance, illumination and noise directly without any prior training or reference image, which distinguishes it from existing supervised-learning methods greatly. Then, the decomposed illumination map will be enhanced by NISM. Having the restored illumination map, the enhancement can be achieved accordingly. Experiments on several public challenging low-light image datasets reveal that the images enhanced by SID-NISM are more natural and have less unexpected artifacts.

29.Two-Stage Copy-Move Forgery Detection with Self Deep Matching and Proposal SuperGlue ⬇️

Copy-move forgery detection identifies a tampered image by detecting pasted and source regions in the same image. In this paper, we propose a novel two-stage framework specially for copy-move forgery detection. The first stage is a backbone self deep matching network, and the second stage is named as Proposal SuperGlue. In the first stage, atrous convolution and skip matching are incorporated to enrich spatial information and leverage hierarchical features. Spatial attention is built on self-correlation to reinforce the ability to find appearance similar regions. In the second stage, Proposal SuperGlue is proposed to remove false-alarmed regions and remedy incomplete regions. Specifically, a proposal selection strategy is designed to enclose highly suspected regions based on proposal generation and backbone score maps. Then, pairwise matching is conducted among candidate proposals by deep learning based keypoint extraction and matching, i.e., SuperPoint and SuperGlue. Integrated score map generation and refinement methods are designed to integrate results of both stages and obtain optimized results. Our two-stage framework unifies end-to-end deep matching and keypoint matching by obtaining highly suspected proposals, and opens a new gate for deep learning research in copy-move forgery detection. Experiments on publicly available datasets demonstrate the effectiveness of our two-stage framework.

30.Domain Adaptive Object Detection via Feature Separation and Alignment ⬇️

Recently, adversarial-based domain adaptive object detection (DAOD) methods have been developed rapidly. However, there are two issues that need to be resolved urgently. Firstly, numerous methods reduce the distributional shifts only by aligning all the feature between the source and target domain, while ignoring the private information of each domain. Secondly, DAOD should consider the feature alignment on object existing regions in images. But redundancy of the region proposals and background noise could reduce the domain transferability. Therefore, we establish a Feature Separation and Alignment Network (FSANet) which consists of a gray-scale feature separation (GSFS) module, a local-global feature alignment (LGFA) module and a region-instance-level alignment (RILA) module. The GSFS module decomposes the distractive/shared information which is useless/useful for detection by a dual-stream framework, to focus on intrinsic feature of objects and resolve the first issue. Then, LGFA and RILA modules reduce the distributional shifts of the multi-level features. Notably, scale-space filtering is exploited to implement adaptive searching for regions to be aligned, and instance-level features in each region are refined to reduce redundancy and noise mentioned in the second issue. Various experiments on multiple benchmark datasets prove that our FSANet achieves better performance on the target domain detection and surpasses the state-of-the-art methods.

31.StrokeGAN: Reducing Mode Collapse in Chinese Font Generation via Stroke Encoding ⬇️

The generation of stylish Chinese fonts is an important problem involved in many applications. Most of existing generation methods are based on the deep generative models, particularly, the generative adversarial networks (GAN) based models. However, these deep generative models may suffer from the mode collapse issue, which significantly degrades the diversity and quality of generated results. In this paper, we introduce a one-bit stroke encoding to capture the key mode information of Chinese characters and then incorporate it into CycleGAN, a popular deep generative model for Chinese font generation. As a result we propose an efficient method called StrokeGAN, mainly motivated by the observation that the stroke encoding contains amount of mode information of Chinese characters. In order to reconstruct the one-bit stroke encoding of the associated generated characters, we introduce a stroke-encoding reconstruction loss imposed on the discriminator. Equipped with such one-bit stroke encoding and stroke-encoding reconstruction loss, the mode collapse issue of CycleGAN can be significantly alleviated, with an improved preservation of strokes and diversity of generated characters. The effectiveness of StrokeGAN is demonstrated by a series of generation tasks over nine datasets with different fonts. The numerical results demonstrate that StrokeGAN generally outperforms the state-of-the-art methods in terms of content and recognition accuracies, as well as certain stroke error, and also generates more realistic characters.

32.Training an Emotion Detection Classifier using Frames from a Mobile Therapeutic Game for Children with Developmental Disorders ⬇️

Automated emotion classification could aid those who struggle to recognize emotion, including children with developmental behavioral conditions such as autism. However, most computer vision emotion models are trained on adult affect and therefore underperform on child faces. In this study, we designed a strategy to gamify the collection and the labeling of child affect data in an effort to boost the performance of automatic child emotion detection to a level closer to what will be needed for translational digital healthcare. We leveraged our therapeutic smartphone game, GuessWhat, which was designed in large part for children with developmental and behavioral conditions, to gamify the secure collection of video data of children expressing a variety of emotions prompted by the game. Through a secure web interface gamifying the human labeling effort, we gathered and labeled 2,155 videos, 39,968 emotion frames, and 106,001 labels on all images. With this drastically expanded pediatric emotion centric database (>30x larger than existing public pediatric affect datasets), we trained a pediatric emotion classification convolutional neural network (CNN) classifier of happy, sad, surprised, fearful, angry, disgust, and neutral expressions in children. The classifier achieved 66.9% balanced accuracy and 67.4% F1-score on the entirety of CAFE as well as 79.1% balanced accuracy and 78.0% F1-score on CAFE Subset A, a subset containing at least 60% human agreement on emotions labels. This performance is at least 10% higher than all previously published classifiers, the best of which reached 56.% balanced accuracy even when combining "anger" and "disgust" into a single class. This work validates that mobile games designed for pediatric therapies can generate high volumes of domain-relevant datasets to train state of the art classifiers to perform tasks highly relevant to precision health efforts.

33.A Closer Look at the Robustness of Vision-and-Language Pre-trained Models ⬇️

Large-scale pre-trained multimodal transformers, such as ViLBERT and UNITER, have propelled the state of the art in vision-and-language (V+L) research to a new level. Although achieving impressive performance on standard tasks, to date, it still remains unclear how robust these pre-trained models are. To investigate, we conduct a host of thorough evaluations on existing pre-trained models over 4 different types of V+L specific model robustness: (i) Linguistic Variation; (ii) Logical Reasoning; (iii) Visual Content Manipulation; and (iv) Answer Distribution Shift. Interestingly, by standard model finetuning, pre-trained V+L models already exhibit better robustness than many task-specific state-of-the-art methods. To further enhance model robustness, we propose Mango, a generic and efficient approach that learns a Multimodal Adversarial Noise GeneratOr in the embedding space to fool pre-trained V+L models. Differing from previous studies focused on one specific type of robustness, Mango is task-agnostic, and enables universal performance lift for pre-trained models over diverse tasks designed to evaluate broad aspects of robustness. Comprehensive experiments demonstrate that Mango achieves new state of the art on 7 out of 9 robustness benchmarks, surpassing existing methods by a significant margin. As the first comprehensive study on V+L robustness, this work puts robustness of pre-trained models into sharper focus, pointing new directions for future study.

34.Automated system to measure Tandem Gait to assess executive functions in children ⬇️

As mobile technologies have become ubiquitous in recent years, computer-based cognitive tests have become more popular and efficient. In this work, we focus on assessing motor function in children by analyzing their gait movements. Although there has been a lot of research on designing automated assessment systems for gait analysis, most of these efforts use obtrusive wearable sensors for measuring body movements. We have devised a computer vision-based assessment system that only requires a camera which makes it easier to employ in school or home environments. A dataset has been created with 27 children performing the test. Furthermore in order to improve the accuracy of the system, a deep learning based model was pre-trained on NTU-RGB+D 120 dataset and then it was fine-tuned on our gait dataset. The results highlight the efficacy of proposed work for automating the assessment of children's performances by achieving 76.61% classification accuracy.

35.Enabling Collaborative Video Sensing at the Edge through Convolutional Sharing ⬇️

While Deep Neural Network (DNN) models have provided remarkable advances in machine vision capabilities, their high computational complexity and model sizes present a formidable roadblock to deployment in AIoT-based sensing applications. In this paper, we propose a novel paradigm by which peer nodes in a network can collaborate to improve their accuracy on person detection, an exemplar machine vision task. The proposed methodology requires no re-training of the DNNs and incurs minimal processing latency as it extracts scene summaries from the collaborators and injects back into DNNs of the reference cameras, on-the-fly. Early results show promise with improvements in recall as high as 10% with a single collaborator, on benchmark datasets.

36.Does the dataset meet your expectations? Explaining sample representation in image data ⬇️

Since the behavior of a neural network model is adversely affected by a lack of diversity in training data, we present a method that identifies and explains such deficiencies. When a dataset is labeled, we note that annotations alone are capable of providing a human interpretable summary of sample diversity. This allows explaining any lack of diversity as the mismatch found when comparing the \textit{actual} distribution of annotations in the dataset with an \textit{expected} distribution of annotations, specified manually to capture essential label diversity. While, in many practical cases, labeling (samples $\rightarrow$ annotations) is expensive, its inverse, simulation (annotations $\rightarrow$ samples) can be cheaper. By mapping the expected distribution of annotations into test samples using parametric simulation, we present a method that explains sample representation using the mismatch in diversity between simulated and collected data. We then apply the method to examine a dataset of geometric shapes to qualitatively and quantitatively explain sample representation in terms of comprehensible aspects such as size, position, and pixel brightness.

37.A grid-point detection method based on U-net for a structured light system ⬇️

Accurate detection of the feature points of the projected pattern plays an extremely important role in one-shot 3D reconstruction systems, especially for the ones using a grid pattern. To solve this problem, this paper proposes a grid-point detection method based on U-net. A specific dataset is designed that includes the images captured with the two-shot imaging method and the ones acquired with the one-shot imaging method. Among them, the images in the first group after labeled as the ground truth images and the images captured at the same pose with the one-shot method are cut into small patches with the size of 64x64 pixels then feed to the training set. The remaining of the images in the second group is the test set. The experimental results show that our method can achieve a better detecting performance with higher accuracy in comparison with the previous methods.

38.Spectral band selection for vegetation properties retrieval using Gaussian processes regression ⬇️

With current and upcoming imaging spectrometers, automated band analysis techniques are needed to enable efficient identification of most informative bands to facilitate optimized processing of spectral data into estimates of biophysical variables. This paper introduces an automated spectral band analysis tool (BAT) based on Gaussian processes regression (GPR) for the spectral analysis of vegetation properties. The GPR-BAT procedure sequentially backwards removes the least contributing band in the regression model for a given variable until only one band is kept. GPR-BAT is implemented within the framework of the free ARTMO's MLRA (machine learning regression algorithms) toolbox, which is dedicated to the transforming of optical remote sensing images into biophysical products. GPR-BAT allows (1) to identify the most informative bands in relating spectral data to a biophysical variable, and (2) to find the least number of bands that preserve optimized accurate predictions. This study concludes that a wise band selection of hyperspectral data is strictly required for optimal vegetation properties mapping.

39.Sparsity-driven Digital Terrain Model Extraction ⬇️

We here introduce an automatic Digital Terrain Model (DTM) extraction method. The proposed sparsity-driven DTM extractor (SD-DTM) takes a high-resolution Digital Surface Model (DSM) as an input and constructs a high-resolution DTM using the variational framework. To obtain an accurate DTM, an iterative approach is proposed for the minimization of the target variational cost function. Accuracy of the SD-DTM is shown in a real-world DSM data set. We show the efficiency and effectiveness of the approach both visually and quantitatively via residual plots in illustrative terrain types.

40.Post-Hurricane Damage Assessment Using Satellite Imagery and Geolocation Features ⬇️

Gaining timely and reliable situation awareness after hazard events such as a hurricane is crucial to emergency managers and first responders. One effective way to achieve that goal is through damage assessment. Recently, disaster researchers have been utilizing imagery captured through satellites or drones to quantify the number of flooded/damaged buildings. In this paper, we propose a mixed data approach, which leverages publicly available satellite imagery and geolocation features of the affected area to identify damaged buildings after a hurricane. The method demonstrated significant improvement from performing a similar task using only imagery features, based on a case study of Hurricane Harvey affecting Greater Houston area in 2017. This result opens door to a wide range of possibilities to unify the advancement in computer vision algorithms such as convolutional neural networks and traditional methods in damage assessment, for example, using flood depth or bare-earth topology. In this work, a creative choice of the geolocation features was made to provide extra information to the imagery features, but it is up to the users to decide which other features can be included to model the physical behavior of the events, depending on their domain knowledge and the type of disaster. The dataset curated in this work is made openly available (DOI: 10.17603/ds2-3cca-f398).

41.Pose Error Reduction for Focus Enhancement in Thermal Synthetic Aperture Visualization ⬇️

Airborne optical sectioning, an effective aerial synthetic aperture imaging technique for revealing artifacts occluded by forests, requires precise measurements of drone poses. In this article we present a new approach for reducing pose estimation errors beyond the possibilities of conventional Perspective-n-Point solutions by considering the underlying optimization as a focusing problem. We present an efficient image integration technique, which also reduces the parameter search space to achieve realistic processing times, and improves the quality of resulting synthetic integral images.

42.FoggySight: A Scheme for Facial Lookup Privacy ⬇️

Advances in deep learning algorithms have enabled better-than-human performance on face recognition tasks. In parallel, private companies have been scraping social media and other public websites that tie photos to identities and have built up large databases of labeled face images. Searches in these databases are now being offered as a service to law enforcement and others and carry a multitude of privacy risks for social media users. In this work, we tackle the problem of providing privacy from such face recognition systems. We propose and evaluate FoggySight, a solution that applies lessons learned from the adversarial examples literature to modify facial photos in a privacy-preserving manner before they are uploaded to social media. FoggySight's core feature is a community protection strategy where users acting as protectors of privacy for others upload decoy photos generated by adversarial machine learning algorithms. We explore different settings for this scheme and find that it does enable protection of facial privacy -- including against a facial recognition service with unknown internals.

43.Attentional Local Contrast Networks for Infrared Small Target Detection ⬇️

To mitigate the issue of minimal intrinsic features for pure data-driven methods, in this paper, we propose a novel model-driven deep network for infrared small target detection, which combines discriminative networks and conventional model-driven methods to make use of both labeled data and the domain knowledge. By designing a feature map cyclic shift scheme, we modularize a conventional local contrast measure method as a depth-wise parameterless nonlinear feature refinement layer in an end-to-end network, which encodes relatively long-range contextual interactions with clear physical interpretability. To highlight and preserve the small target features, we also exploit a bottom-up attentional modulation integrating the smaller scale subtle details of low-level features into high-level features of deeper layers. We conduct detailed ablation studies with varying network depths to empirically verify the effectiveness and efficiency of the design of each component in our network architecture. We also compare the performance of our network against other model-driven methods and deep networks on the open SIRST dataset as well. The results suggest that our network yields a performance boost over its competitors. Our code, trained models, and results are available online.

44.Equalization Loss v2: A New Gradient Balance Approach for Long-tailed Object Detection ⬇️

Recently proposed decoupled training methods emerge as a dominant paradigm for long-tailed object detection. But they require an extra fine-tuning stage, and the disjointed optimization of representation and classifier might lead to suboptimal results. However, end-to-end training methods, like equalization loss (EQL), still perform worse than decoupled training methods. In this paper, we reveal the main issue in long-tailed object detection is the imbalanced gradients between positives and negatives, and find that EQL does not solve it well. To address the problem of imbalanced gradients, we introduce a new version of equalization loss, called equalization loss v2 (EQL v2), a novel gradient guided reweighing mechanism that re-balances the training process for each category independently and equally. Extensive experiments are performed on the challenging LVIS benchmark. EQL v2 outperforms origin EQL by about 4 points overall AP with 14-18 points improvements on the rare categories. More importantly, It also surpasses decoupled training methods. Without further tuning for the Open Images dataset, EQL v2 improves EQL by 6.3 points AP, showing strong generalization ability. Codes will be released at this https URL

45.End-to-end Generative Floor-plan and Layout with Attributes and Relation Graph ⬇️

In this paper, we propose an end-end model for producing furniture layout for interior scene synthesis from the random vector. This proposed model is aimed to support professional interior designers to produce the interior decoration solutions more quickly. The proposed model combines a conditional floor-plan module of the room, a conditional graphical floor-plan module of the room and a conditional layout module. As compared with the prior work on scene synthesis, our proposed three modules enhance the ability of auto-layout generation given the dimensional category of the room. We conduct our experiments on the proposed real-world interior layout dataset that contains $191208$ designs from the professional designers. Our numerical results demonstrate that the proposed model yields higher-quality layouts in comparison with the state-of-the-art model. The dataset and code are released \href{this https URL}{Dataset,Code}

46.Exploration of Whether Skylight Polarization Patterns Contain Three-dimensional Attitude Information ⬇️

Our previous work has demonstrated that Rayleigh model, which is widely used in polarized skylight navigation to describe skylight polarization patterns, does not contain three-dimensional (3D) attitude information [1]. However, it is still necessary to further explore whether the skylight polarization patterns contain 3D attitude information. So, in this paper, a social spider optimization (SSO) method is proposed to estimate three Euler angles, which considers the difference of each pixel among polarization images based on template matching (TM) to make full use of the captured polarization information. In addition, to explore this problem, we not only use angle of polarization (AOP) and degree of polarization (DOP) information, but also the light intensity (LI) information. So, a sky model is established, which combines Berry model and Hosek model to fully describe AOP, DOP, and LI information in the sky, and considers the influence of four neutral points, ground albedo, atmospheric turbidity, and wavelength. The results of simulation show that the SSO algorithm can estimate 3D attitude and the established sky model contains 3D attitude information. However, when there are measurement noise or model error, the accuracy of 3D attitude estimation drops significantly. Especially in field experiment, it is very difficult to estimate 3D attitude. Finally, the results are discussed in detail.

47.Personal Mental Health Navigator: Harnessing the Power of Data, Personal Models, and Health Cybernetics to Promote Psychological Well-being ⬇️

Traditionally, the regime of mental healthcare has followed an episodic psychotherapy model wherein patients seek care from a provider through a prescribed treatment plan developed over multiple provider visits. Recent advances in wearable and mobile technology have generated increased interest in digital mental healthcare that enables individuals to address episodic mental health symptoms. However, these efforts are typically reactive and symptom-focused and do not provide comprehensive, wrap-around, customized treatments that capture an individual's holistic mental health model as it unfolds over time. Recognizing that each individual is unique, we present the notion of Personalized Mental Health Navigation (MHN): a therapist-in-the-loop, cybernetic goal-based system that deploys a continuous cyclic loop of measurement, estimation, guidance, to steer the individual's mental health state towards a healthy zone. We outline the major components of MHN that is premised on the development of an individual's personal mental health state, holistically represented by a high-dimensional cover of multiple knowledge layers such as emotion, biological patterns, sociology, behavior, and cognition. We demonstrate the feasibility of the personalized MHN approach via a 12-month pilot case study for holistic stress management in college students and highlight an instance of a therapist-in-the-loop intervention using MHN for monitoring, estimating, and proactively addressing moderately severe depression over a sustained period of time. We believe MHN paves the way to transform mental healthcare from the current passive, episodic, reactive process (where individuals seek help to address symptoms that have already manifested) to a continuous and navigational paradigm that leverages a personalized model of the individual, promising to deliver timely interventions to individuals in a holistic manner.

48.TEMImageNet and AtomSegNet Deep Learning Training Library and Models for High-Precision Atom Segmentation, Localization, Denoising, and Super-resolution Processing of Atom-Resolution Scanning TEM Images ⬇️

Atom segmentation and localization, noise reduction and super-resolution processing of atomic-resolution scanning transmission electron microscopy (STEM) images with high precision and robustness is a challenging task. Although several conventional algorithms, such has thresholding, edge detection and clustering, can achieve reasonable performance in some predefined sceneries, they tend to fail when interferences from the background are strong and unpredictable. Particularly, for atomic-resolution STEM images, so far there is no well-established algorithm that is robust enough to segment or detect all atomic columns when there is large thickness variation in a recorded image. Herein, we report the development of a training library and a deep learning method that can perform robust and precise atom segmentation, localization, denoising, and super-resolution processing of experimental images. Despite using simulated images as training datasets, the deep-learning model can self-adapt to experimental STEM images and shows outstanding performance in atom detection and localization in challenging contrast conditions and the precision is consistently better than the state-of-the-art two-dimensional Gaussian fit method. Taking a step further, we have deployed our deep-learning models to a desktop app with a graphical user interface and the app is free and open-source. We have also built a TEM ImageNet project website for easy browsing and downloading of the training data.

49.Evaluation of deep learning-based myocardial infarction quantification using Segment CMR software ⬇️

In this paper, the author evaluates the preliminary work related to automating the quantification of the size of the myocardial infarction (MI) using deep learning in Segment cardiovascular magnetic resonance (CMR) software. Here, deep learning is used to automate the segmentation of myocardial boundaries before triggering the automatic quantification of the size of the MI using the expectation-maximization, weighted intensity, a priori information (EWA) algorithm incorporated in the Segment CMR software. Experimental evaluation of the size of the MI shows that more than 50 % (average infarct scar volume), 75% (average infarct scar percentage), and 65 % (average microvascular obstruction percentage) of the network-based results are approximately very close to the expert delineation-based results. Also, in an experiment involving the visualization of myocardial and infarct contours, in all images of the selected stack, the network and expert-based results tie in terms of the number of infarcted and contoured images.

50.PGMAN: An Unsupervised Generative Multi-adversarial Network for Pan-sharpening ⬇️

Pan-sharpening aims at fusing a low-resolution (LR) multi-spectral (MS) image and a high-resolution (HR) panchromatic (PAN) image acquired by a satellite to generate an HR MS image. Many deep learning based methods have been developed in the past few years. However, since there are no intended HR MS images as references for learning, almost all of the existing methods down-sample the MS and PAN images and regard the original MS images as targets to form a supervised setting for training. These methods may perform well on the down-scaled images, however, they generalize poorly to the full-resolution images. To conquer this problem, we design an unsupervised framework that is able to learn directly from the full-resolution images without any preprocessing. The model is built based on a novel generative multi-adversarial network. We use a two-stream generator to extract the modality-specific features from the PAN and MS images, respectively, and develop a dual-discriminator to preserve the spectral and spatial information of the inputs when performing fusion. Furthermore, a novel loss function is introduced to facilitate training under the unsupervised setting. Experiments and comparisons with other state-of-the-art methods on GaoFen-2 and QuickBird images demonstrate that the proposed method can obtain much better fusion results on the full-resolution images.

51.A Differential Model of the Complex Cell ⬇️

The receptive fields of simple cells in the visual cortex can be understood as linear filters. These filters can be modelled by Gabor functions, or by Gaussian derivatives. Gabor functions can also be combined in an `energy model' of the complex cell response. This paper proposes an alternative model of the complex cell, based on Gaussian derivatives. It is most important to account for the insensitivity of the complex response to small shifts of the image. The new model uses a linear combination of the first few derivative filters, at a single position, to approximate the first derivative filter, at a series of adjacent positions. The maximum response, over all positions, gives a signal that is insensitive to small shifts of the image. This model, unlike previous approaches, is based on the scale space theory of visual processing. In particular, the complex cell is built from filters that respond to the \twod\ differential structure of the image. The computational aspects of the new model are studied in one and two dimensions, using the steerability of the Gaussian derivatives. The response of the model to basic images, such as edges and gratings, is derived formally. The response to natural images is also evaluated, using statistical measures of shift insensitivity. The relevance of the new model to the cortical image representation is discussed.

52.Learning-Based Algorithms for Vessel Tracking: A Review ⬇️

Developing efficient vessel-tracking algorithms is crucial for imaging-based diagnosis and treatment of vascular diseases. Vessel tracking aims to solve recognition problems such as key (seed) point detection, centerline extraction, and vascular segmentation. Extensive image-processing techniques have been developed to overcome the problems of vessel tracking that are mainly attributed to the complex morphologies of vessels and image characteristics of angiography. This paper presents a literature review on vessel-tracking methods, focusing on machine-learning-based methods. First, the conventional machine-learning-based algorithms are reviewed, and then, a general survey of deep-learning-based frameworks is provided. On the basis of the reviewed methods, the evaluation issues are introduced. The paper is concluded with discussions about the remaining exigencies and future research.

53.Secret Key Agreement with Physical Unclonable Functions: An Optimality Summary ⬇️

We address security and privacy problems for digital devices and biometrics from an information-theoretic optimality perspective, where a secret key is generated for authentication, identification, message encryption/decryption, or secure computations. A physical unclonable function (PUF) is a promising solution for local security in digital devices and this review gives the most relevant summary for information theorists, coding theorists, and signal processing community members who are interested in optimal PUF constructions. Low-complexity signal processing methods such as transform coding that are developed to make the information-theoretic analysis tractable are discussed. The optimal trade-offs between the secret-key, privacy-leakage, and storage rates for multiple PUF measurements are given. Proposed optimal code constructions that jointly design the vector quantizer and error-correction code parameters are listed. These constructions include modern and algebraic codes such as polar codes and convolutional codes, both of which can achieve small block-error probabilities at short block lengths, corresponding to a small number of PUF circuits. Open problems in the PUF literature from a signal processing, information theory, coding theory, and hardware complexity perspectives and their combinations are listed to stimulate further advancements in the research on local privacy and security.

54.Distilling Optimal Neural Networks: Rapid Search in Diverse Spaces ⬇️

This work presents DONNA (Distilling Optimal Neural Network Architectures), a novel pipeline for rapid neural architecture search and search space exploration, targeting multiple different hardware platforms and user scenarios. In DONNA, a search consists of three phases. First, an accuracy predictor is built for a diverse search space using blockwise knowledge distillation. This predictor enables searching across diverse macro-architectural network parameters such as layer types, attention mechanisms, and channel widths, as well as across micro-architectural parameters such as block repeats, kernel sizes, and expansion rates. Second, a rapid evolutionary search phase finds a Pareto-optimal set of architectures in terms of accuracy and latency for any scenario using the predictor and on-device measurements. Third, Pareto-optimal models can be quickly finetuned to full accuracy. With this approach, DONNA finds architectures that outperform the state of the art. In ImageNet classification, architectures found by DONNA are 20% faster than EfficientNet-B0 and MobileNetV2 on a Nvidia V100 GPU at similar accuracy and 10% faster with 0.5% higher accuracy than MobileNetV2-1.4x on a Samsung S20 smartphone. In addition to neural architecture search, DONNA is used for search-space exploration and hardware-aware model compression.

55.Learning to Run with Potential-Based Reward Shaping and Demonstrations from Video Data ⬇️

Learning to produce efficient movement behaviour for humanoid robots from scratch is a hard problem, as has been illustrated by the "Learning to run" competition at NIPS 2017. The goal of this competition was to train a two-legged model of a humanoid body to run in a simulated race course with maximum speed. All submissions took a tabula rasa approach to reinforcement learning (RL) and were able to produce relatively fast, but not optimal running behaviour. In this paper, we demonstrate how data from videos of human running (e.g. taken from YouTube) can be used to shape the reward of the humanoid learning agent to speed up the learning and produce a better result. Specifically, we are using the positions of key body parts at regular time intervals to define a potential function for potential-based reward shaping (PBRS). Since PBRS does not change the optimal policy, this approach allows the RL agent to overcome sub-optimalities in the human movements that are shown in the videos.
We present experiments in which we combine selected techniques from the top ten approaches from the NIPS competition with further optimizations to create an high-performing agent as a baseline. We then demonstrate how video-based reward shaping improves the performance further, resulting in an RL agent that runs twice as fast as the baseline in 12 hours of training. We furthermore show that our approach can overcome sub-optimal running behaviour in videos, with the learned policy significantly outperforming that of the running agent from the video.

56.Cross-Cohort Generalizability of Deep and Conventional Machine Learning for MRI-based Diagnosis and Prediction of Alzheimer's Disease ⬇️

This work validates the generalizability of MRI-based classification of Alzheimer's disease (AD) patients and controls (CN) to an external data set and to the task of prediction of conversion to AD in individuals with mild cognitive impairment (MCI). We used a conventional support vector machine (SVM) and a deep convolutional neural network (CNN) approach based on structural MRI scans that underwent either minimal pre-processing or more extensive pre-processing into modulated gray matter (GM) maps. Classifiers were optimized and evaluated using cross-validation in the ADNI (334 AD, 520 CN). Trained classifiers were subsequently applied to predict conversion to AD in ADNI MCI patients (231 converters, 628 non-converters) and in the independent Health-RI Parelsnoer data set. From this multi-center study representing a tertiary memory clinic population, we included 199 AD patients, 139 participants with subjective cognitive decline, 48 MCI patients converting to dementia, and 91 MCI patients who did not convert to dementia. AD-CN classification based on modulated GM maps resulted in a similar AUC for SVM (0.940) and CNN (0.933). Application to conversion prediction in MCI yielded significantly higher performance for SVM (0.756) than for CNN (0.742). In external validation, performance was slightly decreased. For AD-CN, it again gave similar AUCs for SVM (0.896) and CNN (0.876). For prediction in MCI, performances decreased for both SVM (0.665) and CNN (0.702). Both with SVM and CNN, classification based on modulated GM maps significantly outperformed classification based on minimally processed images. Deep and conventional classifiers performed equally well for AD classification and their performance decreased only slightly when applied to the external cohort. We expect that this work on external validation contributes towards translation of machine learning to clinical practice.

57.Learning-Based Quality Assessment for Image Super-Resolution ⬇️

Image Super-Resolution (SR) techniques improve visual quality by enhancing the spatial resolution of images. Quality evaluation metrics play a critical role in comparing and optimizing SR algorithms, but current metrics achieve only limited success, largely due to the lack of large-scale quality databases, which are essential for learning accurate and robust SR quality metrics. In this work, we first build a large-scale SR image database using a novel semi-automatic labeling approach, which allows us to label a large number of images with manageable human workload. The resulting SR Image quality database with Semi-Automatic Ratings (SISAR), so far the largest of SR-IQA database, contains 8,400 images of 100 natural scenes. We train an end-to-end Deep Image SR Quality (DISQ) model by employing two-stream Deep Neural Networks (DNNs) for feature extraction, followed by a feature fusion network for quality prediction. Experimental results demonstrate that the proposed method outperforms state-of-the-art metrics and achieves promising generalization performance in cross-database tests. The SISAR database and DISQ model will be made publicly available to facilitate reproducible research.

58.Responsible Disclosure of Generative Models Using Scalable Fingerprinting ⬇️

Over the past five years, deep generative models have achieved a qualitative new level of performance. Generated data has become difficult, if not impossible, to be distinguished from real data. While there are plenty of use cases that benefit from this technology, there are also strong concerns on how this new technology can be misused to spoof sensors, generate deep fakes, and enable misinformation at scale. Unfortunately, current deep fake detection methods are not sustainable, as the gap between real and fake continues to close. In contrast, our work enables a responsible disclosure of such state-of-the-art generative models, that allows researchers and companies to fingerprint their models, so that the generated samples containing a fingerprint can be accurately detected and attributed to a source. Our technique achieves this by an efficient and scalable ad-hoc generation of a large population of models with distinct fingerprints. Our recommended operation point uses a 128-bit fingerprint which in principle results in more than $10^{36}$ identifiable models. Experimental results show that our method fulfills key properties of a fingerprinting mechanism and achieves effectiveness in deep fake detection and attribution.

59.Wasserstein Contrastive Representation Distillation ⬇️

The primary goal of knowledge distillation (KD) is to encapsulate the information of a model learned from a teacher network into a student network, with the latter being more compact than the former. Existing work, e.g., using Kullback-Leibler divergence for distillation, may fail to capture important structural knowledge in the teacher network and often lacks the ability for feature generalization, particularly in situations when teacher and student are built to address different classification tasks. We propose Wasserstein Contrastive Representation Distillation (WCoRD), which leverages both primal and dual forms of Wasserstein distance for KD. The dual form is used for global knowledge transfer, yielding a contrastive learning objective that maximizes the lower bound of mutual information between the teacher and the student networks. The primal form is used for local contrastive knowledge transfer within a mini-batch, effectively matching the distributions of features between the teacher and the student networks. Experiments demonstrate that the proposed WCoRD method outperforms state-of-the-art approaches on privileged information distillation, model compression and cross-modal transfer.

60.Mitigating bias in calibration error estimation ⬇️

Building reliable machine learning systems requires that we correctly understand their level of confidence. Calibration focuses on measuring the degree of accuracy in a model's confidence and most research in calibration focuses on techniques to improve an empirical estimate of calibration error, ECE_bin. Using simulation, we show that ECE_bin can systematically underestimate or overestimate the true calibration error depending on the nature of model miscalibration, the size of the evaluation data set, and the number of bins. Critically, ECE_bin is more strongly biased for perfectly calibrated models. We propose a simple alternative calibration error metric, ECE_sweep, in which the number of bins is chosen to be as large as possible while preserving monotonicity in the calibration function. Evaluating our measure on distributions fit to neural network confidence scores on CIFAR-10, CIFAR-100, and ImageNet, we show that ECE_sweep produces a less biased estimator of calibration error and therefore should be used by any researcher wishing to evaluate the calibration of models trained on similar datasets.

61.CUDA-Optimized real-time rendering of a Foveated Visual System ⬇️

The spatially-varying field of the human visual system has recently received a resurgence of interest with the development of virtual reality (VR) and neural networks. The computational demands of high resolution rendering desired for VR can be offset by savings in the periphery, while neural networks trained with foveated input have shown perceptual gains in i.i.d and o.o.d generalization. In this paper, we present a technique that exploits the CUDA GPU architecture to efficiently generate Gaussian-based foveated images at high definition (1920x1080 px) in real-time (165 Hz), with a larger number of pooling regions than previous Gaussian-based foveation algorithms by several orders of magnitude, producing a smoothly foveated image that requires no further blending or stitching, and that can be well fit for any contrast sensitivity function. The approach described can be adapted from Gaussian blurring to any eccentricity-dependent image processing and our algorithm can meet demand for experimentation to evaluate the role of spatially-varying processing across biological and artificial agents, so that foveation can be added easily on top of existing systems rather than forcing their redesign (emulated foveated renderer). Altogether, this paper demonstrates how a GPU, with a CUDA block-wise architecture, can be employed for radially-variant rendering, with opportunities for more complex post-processing to ensure a metameric foveation scheme. Code is provided.

62.An anatomically-informed 3D CNN for brain aneurysm classification with weak labels ⬇️

A commonly adopted approach to carry out detection tasks in medical imaging is to rely on an initial segmentation. However, this approach strongly depends on voxel-wise annotations which are repetitive and time-consuming to draw for medical experts. An interesting alternative to voxel-wise masks are so-called "weak" labels: these can either be coarse or oversized annotations that are less precise, but noticeably faster to create. In this work, we address the task of brain aneurysm detection as a patch-wise binary classification with weak labels, in contrast to related studies that rather use supervised segmentation methods and voxel-wise delineations. Our approach comes with the non-trivial challenge of the data set creation: as for most focal diseases, anomalous patches (with aneurysm) are outnumbered by those showing no anomaly, and the two classes usually have different spatial distributions. To tackle this frequent scenario of inherently imbalanced, spatially skewed data sets, we propose a novel, anatomically-driven approach by using a multi-scale and multi-input 3D Convolutional Neural Network (CNN). We apply our model to 214 subjects (83 patients, 131 controls) who underwent Time-Of-Flight Magnetic Resonance Angiography (TOF-MRA) and presented a total of 111 unruptured cerebral aneurysms. We compare two strategies for negative patch sampling that have an increasing level of difficulty for the network and we show how this choice can strongly affect the results. To assess whether the added spatial information helps improving performances, we compare our anatomically-informed CNN with a baseline, spatially-agnostic CNN. When considering the more realistic and challenging scenario including vessel-like negative patches, the former model attains the highest classification results (accuracy$\simeq$95%, AUROC$\simeq$0.95, AUPR$\simeq$0.71), thus outperforming the baseline.

63.Jet tagging in the Lund plane with graph networks ⬇️

The identification of boosted heavy particles such as top quarks or vector bosons is one of the key problems arising in experimental studies at the Large Hadron Collider. In this article, we introduce LundNet, a novel jet tagging method which relies on graph neural networks and an efficient description of the radiation patterns within a jet to optimally disentangle signatures of boosted objects from background events. We apply this framework to a number of different benchmarks, showing significantly improved performance for top tagging compared to existing state-of-the-art algorithms. We study the robustness of the LundNet taggers to non-perturbative and detector effects, and show how kinematic cuts in the Lund plane can mitigate overfitting of the neural network to model-dependent contributions. Finally, we consider the computational complexity of this method and its scaling as a function of kinematic Lund plane cuts, showing an order of magnitude improvement in speed over previous graph-based taggers.