Search

Capturing long-range dependencies during feature extraction is crucial for video-based person re-identification (re-id) since it would help to tackle many challenging problems such as occlusion and dramatic pose variation. Moreover, capturing subtle differences, such as bags and glasses, is indispensable to distinguish similar pedestrians. In this paper, we propose a novel and efficacious Statistic Attention (SA) block which can capture both the long-range dependencies and subtle differences. SA block leverages high-order statistics of feature maps, which contain both long-range and high-order information. By modeling relations with these statistics, SA block can explicitly capture long-range dependencies with less time complexity. In addition, high-order statistics usually concentrate on details of feature maps and can perceive the subtle differences between pedestrians. In this way, SA block is capable of discriminating pedestrians with subtle differences. Furthermore, this lightweight block can be conveniently inserted into existing deep neural networks at any depth to form Statistic Attention Network (SANet). To evaluate its performance, we conduct extensive experiments on two challenging video re-id datasets, showing that our SANet outperforms the state-of-the-art methods. Furthermore, to show the generalizability of SANet, we evaluate it on three image re-id datasets and two more general image classification datasets, including ImageNet. The source code is available at http://vipl.ict.ac.cn/resources/codes/code/SANet_code.zip. [ABSTRACT FROM AUTHOR]

In real-world applications, complete or incomplete multi-view data are common, which leads to the problem of generalized multi-view clustering. Recently, researchers attempt to learn the latent representation in the common subspace from heterogeneous data, which usually suffers from feature degeneration. Moreover, there are limited efforts on simultaneously revealing the underlying subspace structure and exploring the complementary information from incomplete multiple views. In this paper, we introduce a novel Generalized Multi-view Collaborative Subspace Clustering (GMCSC) framework to address the above issues, in which consensus subspace structure of all views and embedding subspaces for each view are jointly learned to benefit each other. Specifically, we develop a novel collaborative subspace learning strategy based on self-representation learning, which provides a brand-new way of pursuing the complete subspace structure directly from multi-view data. Furthermore, we explore complementary information by enforcing the consistency across different views and preserving the view-specific information of each view, which can alleviate the problem of feature degeneration and enhance the reasonability of using a consensus representation for multiple views. Experimental results on six benchmark datasets demonstrate that the proposed method can significantly outperform the state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

Decentralized learning has gained great popularity to improve learning efficiency and preserve data privacy. Each computing node makes equal contribution to collaboratively learn a Deep Learning model. The elimination of centralized Parameter Servers (PS) can effectively address many issues such as privacy, performance bottleneck and single-point-failure. However, how to achieve Byzantine Fault Tolerance in decentralized learning systems is rarely explored, although this problem has been extensively studied in centralized systems. In this paper, we present an in-depth study towards the Byzantine resilience of decentralized learning systems with two contributions. First, from the adversarial perspective, we theoretically illustrate that Byzantine attacks are more dangerous and feasible in decentralized learning systems: even one malicious participant can arbitrarily alter the models of other participants by sending carefully crafted updates to its neighbors. Second, from the defense perspective, we propose Ubar, a novel algorithm to enhance decentralized learning with Byzantine Fault Tolerance. Specifically, Ubar provides a Uniform Byzantine-resilient Aggregation Rule for benign nodes to select the useful parameter updates and filter out the malicious ones in each training iteration. It guarantees that each benign node in a decentralized system can train a correct model under very strong Byzantine attacks with an arbitrary number of faulty nodes. We conduct extensive experiments on standard image classification tasks and the results indicate that Ubar can effectively defeat both simple and sophisticated Byzantine attacks with higher performance efficiency than existing solutions. [ABSTRACT FROM AUTHOR]

Recent works on learned image compression perform encoding and decoding processes in a full-resolution manner, resulting in two problems when deployed for practical applications. First, parallel acceleration of the autoregressive entropy model cannot be achieved due to serial decoding. Second, full-resolution inference often causes the out-of-memory (OOM) problem with limited GPU resources, especially for high-resolution images. Block partition is a good choice to handle the above issues, but it brings about new challenges in reducing the redundancy between blocks and eliminating block effects. To tackle the above challenges, this paper provides a learned block-based hybrid image compression (LBHIC) framework. Specifically, we introduce explicit intra prediction into a learned image compression framework to utilize the relation among adjacent blocks. Superior to context modeling by linear weighting of neighbor pixels in traditional codecs, we propose a contextual prediction module (CPM) to better capture long-range correlations by utilizing the strip pooling to extract the most relevant information in neighboring latent space, thus achieving effective information prediction. Moreover, to alleviate blocking artifacts, we further propose a boundary-aware postprocessing module (BPM) with the edge importance taken into account. Extensive experiments demonstrate that the proposed LBHIC codec outperforms the VVC, with a bit-rate conservation of 4.1%, and reduces the decoding time by approximately 86.7% compared with that of state-of-the-art learned image compression methods. [ABSTRACT FROM AUTHOR]

In this paper, we propose an iterative feature matching framework for self-supervised depth estimation in indoor scenes. Conventional methods usually leverage the structure-from-motion supervision to help the photometric optimization escape from the local minima, which have complex ego-motion and large regions with non-texture or repeated-texture. However, the supervision is limited as the reconstruction is usually sparse. To address this, we propose an iterative feature matching framework called IFMNet to jointly learn depths and search for correspondences. With the predicted depths from the previous iteration, we present an online optimized grid searching algorithm to find more accurate correspondences. Given these new correspondences, we compute the triangulated depths and improve the depth network with adaptive bin-wise online hard example mining. Experimental results on the NYU Depth V2 and SceneNet datasets verify the effectiveness of our approach. [ABSTRACT FROM AUTHOR]

Video inpainting is a task of synthesizing spatio-temporal coherent content in missing regions of the given video sequence, which has recently drawn increasing attention. To utilize the temporal information across frames, most recent deep learning-based methods align reference frames to target frame firstly with explicit or implicit motion estimation and then integrate the information from the aligned frames. However, their performance relies heavily on the accuracy of frame-to-frame alignment. To alleviate the above problem, in this paper, a novel Temporal Group Fusion Network (TGF-Net) is proposed to effectively integrate temporal information through a two-stage fusion strategy. Specifically, the input frames are reorganized into different groups, where each group is followed by an intra-group fusion module to integrate information within the group. Different groups provide complementary information for the missing region. A temporal attention model is further designed to adaptively integrate the information across groups. Such a temporal information fusion way gets rid of the dependence on alignment operations, greatly improving the visual quality and temporal consistency of the inpainted results. In addition, a coarse alignment model is introduced at the beginning of the network to handle videos with large motion. Extensive experiments on DAVIS and Youtube-VOS datasets demonstrate the superiority of our proposed method in terms of PSNR/SSIM values, visual quality and temporal consistency, respectively. [ABSTRACT FROM AUTHOR]

The deep learning based methods have improved the visual tracking precision significantly. However, the background distraction and the high precise localization remain challenging problems. Despite that some methods have fused the deep and shallow layer features to solve these problems, the existing fusion methods, like simply concatenating or adding the features from the different layers, cannot take the advantage of both the deep and shallow layer features fully. In this paper, we propose a new adaptive feature fusion method, called the instance-based feature pyramid (IBFP) to obtain the discriminative high-resolution feature, which not only inherits the discriminative information from the deep layer feature, but also keeps the high precision localization information of the shallow layer feature. For utilizing the deep and shallow features effectively, we design an instance-based upsampling (IBU) module to fuse them, and a compressed space channel selection (CSCS) module to re-weight the feature channels adaptively. We insert the IBU and CSCS modules in the Siamese tracker for end-to-end training and testing. By using the proposed IBU and CSCS modules, we fuse the deep and shallow features in a series manner. Experiments on large-scale benchmark datasets demonstrate that the proposed modules boost the capabilities of distinguishing the targets and the similar distractors and perform favorably against the state-of-the-art. [ABSTRACT FROM AUTHOR]

As an important perceptual characteristic of the Human Visual System (HVS), the Just Noticeable Difference (JND) has been studied for decades with image and video processing (e.g., perceptual visual signal compression). However, there is little exploration on the existence of JND for the Deep Machine Vision (DMV), although the DMV has made great strides in many machine vision tasks. In this paper, we take an initial attempt, and demonstrate that the DMV has the JND, termed as the DMV-JND. We then propose a JND model for the image classification task in the DMV. It has been discovered that the DMV can tolerate distorted images with average PSNR of only 9.56dB (the lower the better), by generating JND via unsupervised learning with the proposed DMV-JND-NET. In particular, a semantic-guided redundancy assessment strategy is designed to restrain the magnitude and spatial distribution of the DMV-JND. Experimental results on image classification demonstrate that we successfully find the JND for deep machine vision. Our DMV-JND facilitates a possible direction for DMV-oriented image and video compression, watermarking, quality assessment, deep neural network security, and so on. [ABSTRACT FROM AUTHOR]

The scalable extension of High Efficiency Video Coding, SHVC can code Ultra High-Definition (UHD) video, including 360-degree video for various devices to serve a single bitstream with different display resolutions and qualities. To improve the SHVC compression efficiency, this paper proposes a novel intra and inter-frame coding scheme by first separating the common/visually important information and then applying cuboid-based variable size block partitioning and coding process for the common/visually important information in the base layer. In cuboid-based partitioning a video frame is partitioned into arbitrary shaped rectangular regions, known as cuboids, based on the distribution of relatively homogeneous pixel values. As the cuboid adopts a variable block partitioning based on the homogeneity of the data value, the partitioned blocks have better alignment with the object boundary. Moreover, in the cuboid coding process, only the partitioning tree information and a single value for each block need to be coded which takes lower number of bits and computational time compared to the traditional SHVC base layer. To verify the performance of the proposed method we embedded the proposed scheme as a base layer into the standard SHVC reference software and used several popular UHD/360-degree videos. The experimental results indicate that the proposed scalable coding strategy achieves an average of 14.04% BD-Rate reduction and 0.61 dB BD-PSNR gain for UHD/360-video compared to the operation points provided by an SHVC conforming encoder. [ABSTRACT FROM AUTHOR]

With the evolution of electronic devices, such as 3D cameras, addressing the challenges of text localization in 3D video (e.g., for indexing) is increasingly drawing the attention of the multimedia and video processing community. Existing methods focus on 2D video and their performance in the presence of the challenges in 3D video, such as shadow areas associated with text and irregularly sized and shaped text, degrades. This paper proposes the first approach that successfully addresses the challenges of 3D video in addition to those of 2D. It employs a number of innovations, among which, the first is the Generalized Gradient Vector Flow (GGVF) for dominant points detection. The second is the Wavefront concept for text candidate point detection from those dominant points. In addition, an Adaptive B-Spline Polygon Curve Network (ABS-Net) is proposed for accurate text localization in 3D videos by constructing tight fitting bounding polygons using text candidate points. Extensive experiments on custom (3D video) and standard datasets (2D video and scene text) show that the proposed method is practical and useful, and overall outperforms existing state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Single image haze removal, which is to recover the clear version of a hazy image, is a challenging task in computer vision. In this paper, an additive haze model is proposed to approximate the hazy image formation process. In contrast with the traditional optical model, it regards the haze as an additive layer to a clean image. The model thus avoids estimating the medium transmission rate and the global atmospherical light. In addition, based on a critical observation that haze changes gradually and smoothly across the image, a haze smoothness prior is proposed to constrain this model. This prior assumes that the haze layer is much smoother than the clear image. Benefiting from this prior, we can directly separate the clean image from a single hazy image. Experimental results and comparisons with synthetic images and real-world images demonstrate that the proposed method outperforms state-of-the-art single image haze removal algorithms. [ABSTRACT FROM AUTHOR]

Self-supervised representation learning for videos has been very attractive recently because these methods exploit the information inherently obtained from the video itself instead of annotated labels that is quite time-consuming. However, existing methods ignore the importance of global observation while performing spatio-temporal transformation perception, which highly limits the expression capabilities of the video representation. This paper proposes a novel pretext task that combines the temporal information perception of the video with the motion amplitude perception of moving objects to learn the spatio-temporal representation of the video. Specifically, given a video clip containing several video segments, each video segment is sampled by different sampling rates and the order of video segments is disrupted. Then, the network is used to regress the sampling rate of each video segment and classify the order of input video segments. In the pre-training stage, the network can learn rich spatio-temporal semantic information where content-related contrastive learning is introduced to make the learned video representation more discriminative. To alleviate the appearance dependency caused by contrastive learning, we design a novel and robust vector similarity measurement approach, which can take feature alignment into consideration. Moreover, a view synthesis framework is proposed to further improve the performance of contrastive learning by automatically generating reasonable transformed views. We conduct benchmark experiments with several 3D backbone networks on two datasets. The results show that our proposed method outperforms the existing state-of-the-art methods across the three backbones on two downstream tasks of human action recognition and video retrieval. [ABSTRACT FROM AUTHOR]

Video highlight detection (VHD) has emerged as a pressing task due to the unprecedentedly increasing amount of video data, such as those from e-commerce live-broadcasting platforms. Many approaches focus on exploiting text data, in the form of video description or time-sync comments, to facilitate the VHD task. Despite the promising results, they have largely overlooked the noises inherent in the text data and have mostly relied on isolating the feature of text and video. In this paper, we introduce a novel model to handle VHD, termed Coevolution Network (CoEvo-Net), that allows us to account for joint learning of the language and video features explicitly via a coevolution paradigm, in which features from the two data modalities progressively refine each other. This is achieved by a dedicated CoEvo-Cell that takes language and video together as inputs, extracts cross-modality, and filters the undesired parts of the input, such as words in a sentence. Furthermore, we release a large-scale dataset of e-commerce for VHD, in which each video is coupled with a sentence for description, to benchmark the sentence-based VHD approaches. Extensive experiments on the released dataset demonstrate that CoEvo-Net achieves state-of-the-art performance. Our dataset and code will be made publicly available. [ABSTRACT FROM AUTHOR]

A noise-suppressing low-light image enhancement approach is proposed in this paper based on the extent of exposedness at each image pixel. To this end, a progressive, structure-aware exposedness estimation procedure is presented that quantifies local and global exposedness. These exposedness values are leveraged to produce a locally smooth pixel-level map that signifies the required degrees of enhancement at image pixels. This map is subsequently used in an enhancement function, which satisfies a few important properties, to generate the enhanced image. Before the enhancement, inherent noise in the low-light image is diminished employing a detail-preserving, low gradient magnitude suppression method. Subjective and quantitative analysis of results on a wide variety of natural and synthetically generated low-light images from standard databases using PSNR, iRSE, SSIM, and measures of perceptual quality, natural image statistics and brightness preservation suggests that our approach in general outperforms the state-of-the-art. Ablation studies and further experiments show the importance of a few components of our approach, and that our approach is computationally fast. [ABSTRACT FROM AUTHOR]

Numerous image retargeting algorithms have been proposed to achieve adaptive image resizing during the past years. To compare different image retargeting algorithms, reliable objective image retargeting quality assessment (IRQA) metrics are highly desired. Given that image retargeting usually introduces geometric distortions, this paper presents an objective IRQA metric by measuring both local and global geometric distortions (LGGD). Since human visual system perception is highly dependent on edges and the geometric distortions caused by image retargeting usually cause edge deformation, a sketch token-based local edge descriptor (ST-LED) is introduced to represent geometric-aware features in LGGD. First, ST-LED is first applied on both source and retargeted images for edge pattern representation. Second, pixel-level backward registration is conducted to enable estimating local geometric distortion (LGD) and a spatial pyramid-improved Bag-of-Token (BoT) model is built to enable estimating global geometric distortion (GGD). Since the proposed LGGD metric only focuses on geometric distortion while image retargeting quality is related with more aspects, we further fuse LGGD and an existing (EXT) IRQA metric to build a final version called LGGD+ for IRQA. Experiments on two benchmark databases demonstrate the superiority of LGGD+ and the excellent compatibility of our proposed LGGD for further improving a wide range of existing IRQA metrics (including both geometric distortion and non-geometric distortion metrics). In addition, the effectiveness of our LGGD metric is also demonstrated in another relevant task, i.e., quality evaluation of depth-image-based rendering (DIBR)-synthesized images, which also calls for accurate estimation of geometric distortion. [ABSTRACT FROM AUTHOR]

Image captioning is one of the primary goals in computer vision which aims to automatically generate natural descriptions for images. Intuitively, human visual system can notice some stimulating regions at first glance, and then volitionally focus on interesting objects within the region. For example, to generate a free-form sentence about “boy-catch-baseball”, the visual region involving “boy” and “baseball” could be first attended and then guide the salient object discovery for the word-by-word generation. Till now, previous captioning works mainly rely on the object-wise modeling and ignore the rich regional patterns. To mitigate the drawback, this paper proposes the region-aware interaction learning method, which aims to explicitly capture the semantic correlations in the region and object dimensions for the word inference. First, given an image, we extract a set of regions which contain diverse objects and their relations. Second, we present the spatial-GCN interaction refining structure which can establish the connection between regions and objects to effectively capture contextual information. Third, we design the dual-attention interaction inference procedure, which enables attention to be calculated in region and object dimensions jointly for the word generation. Specifically, the guidance mechanism is proposed to selectively emphasize semantic inter-dependencies from region to object attentions. Extensive experiments on the MSCOCO dataset demonstrate the superiority of the proposed method. Additional ablation studies and visualization further validate its effectiveness. [ABSTRACT FROM AUTHOR]

Image classification is a fundamental artificial intelligence task that labels images into one of some predefined classes. However, training complex image classification models requires a large amount of computation resources and data in order to reach state-of-the-art performance. This demand drives the growth of distributed deep learning, where multiple agents cooperatively train global models with their individual datasets. Among such learning systems, decentralized learning is particularly attractive, as it can improve the efficiency and fault tolerance by eliminating the centralized parameter server, which could be the single point of failure or performance bottleneck. Although the agents do not need to disclose their training image samples, they exchange parameters with each other at each iteration, which can put them at the risk of data privacy leakage. Past works demonstrated the possibility of recovering training images from the exchanged parameters. One common defense direction is to adopt Differential Privacy (DP) to secure the optimization algorithms such as Stochastic Gradient Descent (SGD). Those DP-based methods mainly focus on standalone systems, or centralized distributed learning. How to enforce and optimize DP protection in decentralized learning systems is unknown and challenging, due to their complex communication topologies and distinct learning characteristics. In this paper, we design TOP- DP, a novel solution to optimize the differential privacy protection of decentralized image classification systems. The key insight of our solution is to leverage the unique features of decentralized communication topologies to reduce the noise scale and improve the model usability. (1) We enhance the DP-SGD algorithm with this topology-aware noise reduction strategy, and integrate the time-aware noise decay technique. (2) We design two novel learning protocols (synchronous and asynchronous) to protect systems with different network connectivities and topologies. We formally analyze and prove the DP requirement of our proposed solutions. Experimental evaluations demonstrate that our solution achieves a better trade-off between usability and privacy than prior works. To the best of our knowledge, this is the first DP optimization work from the perspective of network topologies. [ABSTRACT FROM AUTHOR]

Human action recognition which recognizes human actions in a video is a fundamental task in computer vision field. Although multiple existing methods with single-view or multi-view have been presented for human action recognition, these recognition approaches cannot be extended into new action recognition or action classification tasks, as well as discover underlying correlations among different views. To tackle the above problem, this paper proposes a new lifelong multi-view subspace learning framework for continuous human action recognition, which could exploit the complementary information amongst different views from a lifelong learning perspective. More specifically, a set of view-specific libraries is established to gradually store the useful information within multiple views. As a new action recognition task comes, we decompose the model parameters into a set of embedded parameters over view-specific libraries. A latent representation subspace is constructed via encouraging it to be close to different view-specific libraries, which can leverage the high-order correlations among different views and further avoid partial information for action recognition task. Meanwhile, we propose to employ an alternating direction strategy to optimize our proposed method. Empirical studies on real-world multi-view action recognition datasets have shown that our proposed framework attains the superior recognition performance and saves the computational time when continually learning new action recognition tasks. [ABSTRACT FROM AUTHOR]

Stereoscopic omnidirectional content, as a novel visual media, has drawn wide attention in recent years due to its ability in providing strong immersive experience. Since Stereoscopic Omnidirectional Images (SOIs) involve the properties from panoramic and stereoscopic visual perception, it is very challenging to establish an efficient and effective visual quality evaluation model for SOIs. To better measure the user’s experience in virtual reality, we put forward a novel deep learning framework to assess the quality of SOIs in this paper. Firstly, the deformable convolutions instead of standard convolutions are adopted to ensure the invariant receptive fields of convolutional kernels on Equi-Rectangular Projection (ERP). Secondly, according to the stereoscopic property, we use binocular-difference information and a coarse-to-fine mechanism to construct the binocular feature extraction network. Thirdly, a three-channel network involving left-view, right-view and binocular-difference channels is presented to simulate the process of monocular and binocular interactions, in which independent quality labels are provided for each channel to reflect the individual effect of monocular and binocular visions on the whole visual quality. Finally, experimental results on two available benchmark databases demonstrate the superiority of the proposed metric over the state-of-the-art blind quality assessment models in predicting the quality of SOIs. Moreover, our model is efficient in computational cost as the feature extraction is directly applied on ERP images. [ABSTRACT FROM AUTHOR]

In the field of frontier research, information security has received a lot of interest, but in the field of information security algorithm, the introduction of decimals makes it impossible to bypass the topic of calculation accuracy. This article creatively proposes the definition and related proofs of double parameters fractal sorting matrix (DPFSM). As a new matrix classification with fractal properties, DPFSM contains self-similar structures in the ordering of both elements and sub-blocks in the matrix. These two self-similar structures are determined by two different parameters. To verify the theory, this paper presents a type of $2\times 2$ DPFSM iterative generation method, as well as the theory, steps, and examples of the iteration. DPFSM is a space position transformation matrix, which has a better periodic law than a single parameter fractal sorting matrix (FSM). The proposal of DPFSM expands the fractal theory and solves the limitation of calculation accuracy on information security. The image encryption algorithm based on DPFSM is proposed, and the security analysis demonstrates the security. DPFSM has good application value in the field of information security. [ABSTRACT FROM AUTHOR]

Most of siamese network and correlation filter (CF) based trackers usually employ the context regression scheme to achieve appealing performance in both accuracy and efficiency. However, they are prone to drifting in challenging situations exhibiting occlusion, out-of-view and large-scale variations due to the lack of failure correction ability in these regressors. Particle filter based trackers can help to recover from tracking failures since several particles of high confidence about the target can be remained for the probability estimation of next frames, but need the large numbers of particles for each frame. In this paper, we propose a generic dynamic particle filter framework, which can reasonably control the number of particles in different scenarios, to improve the robustness of siamese and CF trackers by jointing the target classifier to relieve drifting. Our fundamental insight is that general scenarios are processed efficiently by the context regressor with few particles, while special scenarios are handled effectively by the target classifier with many particles. We propose a novel measure to determine whether to adopt the regressor and few particles to estimate target states with high measure scores, or increase the number of particles to prevent drifting with the proposed multi-template matching strategy in the classifier. In extensive experiments on eight large-scale benchmarks including OTB, UAV, TC128, VOT2017, VOT2019, LaSOT, TrackingNet and Got-10k, the proposed framework enables many basic siamese and CF trackers to operate at least over 24 frames per second and achieve superior tracking performance than themself, as well as the comparable accuracy with the state-of-the-art trackers. Furthermore, our framework is flexible and still has great potential for improvement and generalization. [ABSTRACT FROM AUTHOR]

Labanotation is a symbolic recording system for human movements, and also a powerful tool for protecting and spreading folk dances and other performing arts. State-of-the-art automatic Labanotation uses end-to-end methods with sequence-based skeleton representation, which cannot capture the relationship between joints and bones in the skeleton for accurate descriptions of continuous lower limb movements such as dance steps. In this paper, we propose a novel double-stream fusion method of directed graph neural networks (DGNN), combined with connectionist temporal classification (CTC), namely DFGNN-CTC, for sequential fine-grained motion recognition, such as the Labanotation generation of unsegmented dance movement. First, we extract double-stream directed graph feature, employing an orientation-normalized directed acyclic graph (ON-DAG) and an orientation-normalized temporal directed acyclic graph (ON-TDAG), to jointly model spatiotemporal properties of movement recorded in motion capture data. Then, we design a CTC-based fusion-pooling module to fuse the spatial and temporal streams encoded by two DGNNs. It concatenates and fuses the two streams to generate discriminative descriptions of each time step, and concentrates them to make per-time-step predictions of Laban gesture type, from which the CTC searches the optimal Laban symbol sequence, corresponding to elemental motions composing the movement. In this way, the new method enables much finer discrimination for similar Laban gestures with subtle differences in spatial and temporal properties through joint contextual spatiotemporal modeling so that it achieves much superior performance in continuous Labanotation generation to existing methods, which only have single-stream analysis either spatially or temporally. The experiments on two Labanotation-labelled motion capture datasets demonstrate the effectiveness of the components in the proposed method and its superiority comparing with the state-of-the-art methods, especially for lower limb movements. [ABSTRACT FROM AUTHOR]

Infrared image retains typical thermal targets while visible image preserves rich texture details, image fusion aims to reconstruct a synthesized image containing prominent targets and abundant texture details. Most of deep learning-based methods mainly focus on convolution operation to extract the local features, but do not fully consider their multi-scale characteristics and global dependencies, which may cause loss of target regions and texture details in the fused image. Towards this goal, we present a unified multi-scale densely connected fusion network in this paper, named as UNFusion. We carefully design a multi-scale encoder-decoder architecture that can efficiently extract and reconstruct multi-scale deep features. Dense skip connections are employed in both encoder and decoder sub-networks to reuse all the intermediate features of different layers and scales for fusion tasks. In the fusion layer, $L_{p} $ normalized attention models, which include three kinds of different norms, are proposed to highlight and combine these deep features from spatial and channel dimensions, and the combined spatial and channel attention maps are used to reconstruct a final fused image. We conduct extensive experiments on the public TNO and Roadscene datasets, and the results demonstrate that our UNFusion can simultaneously preserve high brightness of typical thermal targets and abundant texture details to obtain superior scene representation and better visual perception. Besides, our UNFusion achieves better fusion performance and transcends other state-of-the-art methods in terms of qualitative and quantitative comparisons. Our code is available at https://github.com/Zhishe-Wang/UNFusion. [ABSTRACT FROM AUTHOR]

Recently, most coverless image steganography (CIS) methods are based on robust mapping rules. However, due to the limited mapping expression relationship between secret information and hash sequence, it is a challenge to further improve the hiding ability of coverless information hiding. Towards this goal, this paper proposes a robust coverless steganography scheme using camouflage image(CI-CIS). For the sender, CI-CIS introduces an camouflage image as the transmission carrier and establishes the correlation between them by Convolutional Neural Network(CNN) features. For the receiver, the camouflage image can retrieve the corresponding stego-image to recover the secret information. To this end, we designed a reversible retrieval scheme between stego-image and camouflage image by using image clustering. At the same time, since the semantic features represented by CNN are robust to image attacks, our method can increase the capability of the CIS effectively. Besides, we also build an inverted index to improve retrieval efficiency. Experimental results and analysis show that the CI-CIS has higher robustness and more flexible capacity setting compared with the existing CIS methods. [ABSTRACT FROM AUTHOR]

In image processing, smoke may degrade visibility and deteriorate the performance of high-level vision applications. Therefore, single image smoke removal is crucial for computer vision. Currently, existing smoke removal algorithms mainly leverage handcrafted priors. Moreover, these methods usually apply haze removal methods to perform smoke removal due to the similarity between smoke and haze. However, these methods cannot sufficiently address the degradation of thick smoke and may suffer from residual smoke and color distortion problems due to the non-global and non-homogeneous distribution of smoke. In this paper, to solve the aforementioned problems, an end-to-end deep neural network called DesmokeNet is proposed. We construct a two-stage recovered pipeline to remove the smoke in different thicknesses. The light and thick smoke is first removed locally by the smoke removal network (SRN). The missing pixels in the thick smoke are then recovered by the pixel compensation network (PCN). Moreover, we proposed the thickness-aware pixel loss and the dark channel loss to suppress the residual smoke. To further increase the discriminative ability of the DesmokeNet, we proposed self-attentive feature consensus loss and multi-level contrastive regularization loss to improve the performance of smoke removal. Finally, to train the proposed method, we construct the first large-scale dataset containing synthetic and real-world data. Extensive experiments show that the proposed method outperforms favorably against other state-of-the-art methods quantitatively and qualitatively. [ABSTRACT FROM AUTHOR]

Traditional steganography methods are usually designed on a lossless channel; thus, messages are often not extracted correctly from an image transmitted over a lossy channel that includes attacks such as scaling. To address this issue, in recent years, the field of robust steganography has emerged. In this paper, the process of image scaling by interpolation is first observed and serves as the basis for proposing the idea of inverse interpolation. Subsequently, the idea of constructing an inverse interpolation equation set is proposed to solve the problem of intersectional blocks during the inverse interpolation process. Then, the scaling factor’s valid range of inverse interpolation is analyzed. Next, the inverse interpolation is successfully applied in robust image steganography. A method that combines antiscaling and antidetection is proposed. Afterward, actual tests on the top 9 mobile phone brands with 28 models and 2 social communication apps that are currently popular in China are done. The scaling factor’s valid range of the proposed method is verified to match the actual lossy channel. The experimental results show that the proposed method achieves a reliable extraction of embedded messages for common interpolation scaling attacks while maintaining high statistical detection resistance. [ABSTRACT FROM AUTHOR]

Face manipulation techniques improve fast with the development of powerful image generation models. Two particular face manipulation methods, namely face swap and expression reenactment attract much attention for their flexibility and ease to generate high quality synthesis results. Recently, these two subjects are actively studied. However, most existing methods treat the two tasks separately, ignoring their underlying similarity. In this paper, we propose to tackle the two problems within a unified framework that achieves high quality synthesis results. The enabling component for our unified framework is the clean disentanglement of 3D pose, shape, and expression factors and then recombining them for different tasks accordingly. We then use the same set of 2D representations for face swap and expression reenactment tasks that are input to a common image translation model to directly generate the final synthetic images. Once trained, the proposed model can accomplish both face swap and expression reenactment tasks for previously unseen subjects. Comprehensive experiments and comparisons show that the proposed method achieves high fidelity results in multiple aspects, and it is especially good at faithfully preserving source facial shape in the face swap task, and accurately transferring facial movements in the expression reenactment task. [ABSTRACT FROM AUTHOR]

In video coding, it is a challenge to deal with scenes with complex motions, such as rotation and zooming. Although affine motion compensation (AMC) is employed in Versatile Video Coding (VVC), it is still difficult to handle non-translational motions due to the adopted hand-craft block-based motion compensation. In this paper, we propose a deep affine motion compensation network (DAMC-Net) for inter prediction in video coding to effectively improve the prediction accuracy. To the best of our knowledge, our work is the first attempt to deal with the deformable motion compensation based on CNN in VVC. Specifically, a deformable motion-compensated prediction (DMCP) module is proposed to compensate the current encoding block through a learnable way to estimate accurate motion fields. Meanwhile, the spatial neighboring information and the temporal reference block as well as the initial motion field are fully exploited. By effectively fusing the multi-channel feature maps from DMCP, an attention-based fusion and reconstruction (AFR) module is designed to reconstruct the output block. The proposed DAMC-Net is integrated into VVC and the experimental results demonstrate that the proposed method considerably enhances the coding performance. [ABSTRACT FROM AUTHOR]

Alongside classic Visible Light Communication using LEDs and photodiodes, Optical Display-Camera Communication offers another promising way of transmitting data through visible light wirelessly: A display serves as a multi-parallel transmitter array and a camera as the corresponding receiver. Due to the high number of pixels, high data rates in the range of several Mbit/s can be achieved despite comparatively low temporal rates. A major challenge that all high rate system realizations have to overcome is the lack of synchronization between display and camera. Along with varying playback and recording behavior of the devices, this typically results in mixed recordings of display frames in the receiver, which complicates frame decoding. In this paper, we fundamentally investigate the temporal sampling process in display-camera communication. Using our improved modeling we determine the minimum temporal oversampling factor required for lossless frame acquisition, in order to maximize system efficiency in the temporal domain. We propose a solution for the synchronization problem for invisible system approaches with differential modulation, in which the original display frames are recovered from asynchronous recordings in the receiver. Experimental results show that this method works robustly even with high motion background videos and without the need for embedded synchronization markers. [ABSTRACT FROM AUTHOR]

Near-infrared and visual (NIR-VIS) face matching, as the most typical task in Heterogeneous Face Recognition (HFR), has attracted increasing attention in recent years. However, due to the large within-class discrepancies, including domain differences and residual discrepancies (i.e., lighting, expressions, occlusion, blurry, pose, etc), this is still a difficult task. Conventional NIR-VIS FR methods only focus on reducing the modality gap between cross-domain images, while neglecting to eliminate the residual variations. To better solve the above problems, this paper proposes a novel Orthogonal Modality Disentanglement and Representation Alignment (OMDRA) approach, which consists of three key components, including Modality-Invariant (MI) loss, Orthogonal Modality Disentanglement (OMD) and Deep Representation Alignment (DRA). Firstly, the MI loss is designed to learn modality-invariant and identity-discriminative representation, by increasing between-class separability and within-class compactness between NIR and VIS heterogeneous data. Secondly, the high-level Hybrid Facial Feature (HFF) layer of the backbone network is projected into two subspaces: the modality-related and identity-related subspaces. The OMD is designed to decouple modal information via an adversarial process, and we further impose Orthogonal Representation Decorrelation (ORD) to the OMD to decrease the correlation between identity representations and domain representations, as well as enhancing their representation capabilities. Finally, the DRA aims to eliminate the residual variations by performing a high-level representation alignment between non-neutral face and neutral face, which can effectively guides the network to learn discriminative and residual-invariant face representation. The joint scheme enables the disentanglement of modality variations, elimination of residual discrepancies, and the purification of identity information. Extensive experiments on challenging cross-domain databases indicate that our OMDRA method is superior to the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Multi-modal services, typically integrated by visual, audio, and haptic signals, have been considered as promising killer services in 5G and beyond 5G era. However, due to essential difference among these signals, how to guarantee quality of multi-modal services is a significant technical challenge. Existing transmission schemes, which deliver and process each modality signal separately, cannot meet such requirements as low latency, high reliability, and high throughput. To get over the dilemma, this paper proposes a general cross-modal transmission strategy by taking advantage of the potential correlation among modalities, which consists of a delivery mechanism at the sender and a signal restoration procedure at the receiver. On the one hand, a visual-aided haptic content compression method is designed for the delivery mechanism. By utilizing the category correlation among modalities, haptic signals with similar visual content can be effectively compressed, reducing transmission burden. On the other hand, a fine-grained haptic to image synthesis (FHIS) approach is proposed for realizing signal restoration. Through exploring strong matching properties among modalities, the FHIS can restore the impaired, missing, delayed visual images from the received haptic signals. Experiments on a standard visual-haptic database and a practical platform evaluate the performance of the constructed cross-modal transmission strategy. [ABSTRACT FROM AUTHOR]

Real-world scenes always exhibit objects with clutter backgrounds, posing great challenges for deep salient object detection models. In this paper, we propose salient object detection by engaging two saliency cues, i.e., the part-whole hierarchies and contrast cues, resulting in a PWHCNet. Specifically, two branches, which consists of a Dynamic Grouping Capsules (DGC) branch and a DenseHRNet branch, are put in place to learn the part-whole hierarchies and contrast cues, respectively. Moreover, to help highlight the whole salient object in complex scenes, a Background Suppression (BS) module is proposed to guide the shallow features of DenseHRNet with the aid of the part-whole relational cues captured by DGC. Subsequently, these two saliency cues are integrated via a Self-Channel and Mutual-Spatial (SCMS) attention mechanism. Experimental results on five benchmarks demonstrate that the proposed PWHCNet achieves state-of-the-art performance while obtaining the whole salient objects with fine details. [ABSTRACT FROM AUTHOR]

The emergence of wearable devices has opened up new potentials for egocentric activity recognition. Although some methods integrate attention mechanisms into deep neural networks to capture fine-grained human-object interactions in a weak-supervision manner, they either ignore exploiting the temporal consistency or generate attention based on considering appearance cues only. To address these limitations, in this paper, we propose an enhanced attention-tracking method, combined with multi-branch network (EAT-MBNet), for egocentric activity recognition. Specifically, we propose class-aware attention maps (CAAMs) by employing a self-attention-based module to refine the class activation maps (CAMs). Our proposed method can enhance the semantic dependency between the activity categories and the feature maps. To highlight the discriminative features from the regions of interest across frames, we propose a flow-guided attention-tracking (F-AT) module, by simultaneously leveraging historical attention and motion patterns. Furthermore, we propose a cross-modality modeling branch based on an interactive GRU module, which captures the time-synchronized long-term relationships between the appearance and motion branches. Experimental results on four egocentric activity benchmarks demonstrate that the proposed method achieves state-of-the-art performance. [ABSTRACT FROM AUTHOR]

The edge information plays a key role in the restoration of a depth map. Most conventional methods assume that the color image and depth map are consistent in edge areas. However, complex texture regions in the color image do not match exactly with edges in the depth map. In this paper, firstly, we point out that in most cases the consistency between normal map and depth map is much higher than that between RGB-D pairs. Then we propose a dual normal-depth regularization term to guide the restoration of depth map, which constrains the edge consistency between normal map and depth map back and forth. Moreover, considering the bimodal characteristic of weight distribution that exists in depth discontinuous areas, a reweighted graph Laplacian regularizer is proposed to promote this bimodal characteristic. And this regularization is incorporated into a unified optimization framework to effectively protect the piece-wise smoothness(PWS) characteristics of depth map. By treating depth image as graph signal, the weight between two nodes is adapted according to its content. The proposed method is tested for both noise-free and noisy cases, and is compared against the state-of-the-art methods on both synthesis and real captured datasets. Extensive experimental results demonstrate the superior performance of our method compared with most state-of-the-art works in terms of both objective and subjective quality evaluations. Specifically, our method is more effective on edge areas and more robust to noises. [ABSTRACT FROM AUTHOR]

Multimodal magnetic resonance imaging (MRI) provides complementary information about targets, and the segmentation of multimodal MRI is widely used as an essential preprocessing step for initial diagnosis, stage differentiation, and post-treatment efficacy evaluation in clinical situations. For the main modality or each of the modalities, it is important to enhance the visual information by modeling the connection and effectively fusing the features among them. However, the existing methods for multimodal segmentation have a drawback; they coincidentally drop information of individual modality during the fusion process. Recently, graph learning-based methods have been applied in segmentation, and these methods have achieved considerable improvements by modeling the relationships across feature regions and reasoning using global information. In this paper, we propose a graph learning-based approach to efficiently extract modality-specific features and establish regional correspondence effectively among all modalities. In detail, after projecting features into a graph domain and employing graph convolution to propagate information across all regions for learning global modality-specific features, we propose a mutual information-based graph co-attention module to learn the weight coefficients of one bipartite graph constructed by the fully connected graphs having different modalities in the graph domain and by selectively fusing the node features. Based on the deformation diagram between the spatial-graph space and our proposed graph co-attention module, we present a multimodal prior-guided segmentation framework, which uses two strategies for two clinical situations: Modality-Specific Learning Strategy and Co-Modality Learning Strategy. Besides, the improved Co-Modality Learning Strategy is used with trainable weights in the multi-task loss for the optimization of the proposed framework. We validated our proposed modules and frameworks on two multimodal MRI datasets: our private liver lesion dataset and a public prostate zone dataset. Our experimental results on both datasets prove the superiority of our proposed approaches. [ABSTRACT FROM AUTHOR]

Facial Expression Recognition (FER) is a long-standing but challenging research problem in computer vision. Existing approaches mainly focus on single-label emotional prediction, which cannot handle the complex multi-label FER task because of the coupling behavior of multiple emotions on a single facial image. To this end, in this paper, we propose a novel Self-supervised Exclusive-Inclusive Interactive Learning (SEIIL) method to facilitate discriminative multi-label FER in the wild, which can effectively handle the coupled multiple sentiments with limited unconstrained training data. Specifically, we construct an emotion disentangling module to capture the inclusive and exclusive characteristics of facial expressions, which can decouple the compound numerous emotions on an image. Moreover, an adaptively-weighted ensemble technique is conceived to aggregate category-level latent exclusive embeddings, and then a conditional adversarial interactive learning module is designed to fully leverage the complementary between the inclusive and formulated latent representations. Furthermore, to tackle the insufficient data for training, we introduce a self-supervised learning strategy to augment the amount and diversity of facial images, which can endow the model with advanced generalization ability. Under this strategy, the proposed two modules can be concurrently utilized in our SEIIL to jointly handle the coupled emotions and alleviate the overfitting problem. Extensive experimental results on six databases illustrate the superb performance of our method against state-of-the-art baselines. [ABSTRACT FROM AUTHOR]

Temporal action detection is a challenging task in video understanding, due to the complexity of the background and rich action content impacting high-quality temporal proposals generation in untrimmed videos. Capsule networks can avoid some limitations of the invariance caused by pooling and inability from convolutional neural networks, which can better understand the temporal relations for temporal action detection. However, because of the extremely computationally expensive procedure, capsule network is difficult to be applied to the task of temporal action detection. To address this issue, this paper proposes a novel U-shaped capsule network framework with a k-Nearest Neighbor (k-NN) mechanism of 3D convolutional dynamic routing, which we named U-BlockConvCaps. Furthermore, we build a Capsules Boundary Network (CapsBoundNet) based on U-BlockConvCaps for dense temporal action proposal generation. Specifically, the first module is one 1D convolutional layer for fusing the two-stream with RGB and optical flow video features. The sampling module further processes the fused features to generate the 2D start-end action proposal feature maps. Then, the multi-scale U-Block convolutional capsule module with 3D convolutional dynamic routing is used to process the proposal feature map. Finally, the feature maps generated from the CapsBoundNet are used to predict starting, ending, action classification, and action regression score maps, which help to capture the boundary and intersection over union features. Our work innovatively improves the dynamic routing algorithm of capsule networks and extends the use of capsule networks to the temporal action detection task for the first time in the literature. The experimental results on benchmarks THUMOS14 show that the performance of CapsBoundNet is obviously beyond the state-of-the-art methods, e.g., the mAP@tIoU = 0.3, 0.4, 0.5 on THUMOS14 are improved from 63.6% to 70.0%, 57.8% to 63.1%, 51.3% to 52.9%, respectively. We also got competitive results on the action detection dataset of ActivityNet1.3. [ABSTRACT FROM AUTHOR]

Recently, VR sickness assessment for VR videos is highly demanded in industry and research fields to address VR viewing safety issues. Especially, it is difficult to evaluate VR sickness of individuals due to individual differences. To achieve the challenging goal, we focus on deep feature fusion of sickness-related information. In this paper, we propose a novel deep learning-based assessment framework which estimates VR sickness of individual viewers with VR videos and corresponding physiological responses. We design the content stimulus guider imitating the phenomenon that humans feel VR sickness. The content stimulus guider extracts a deep stimulus feature from a VR video to reflect VR sickness caused by VR videos. In addition, we devise the physiological response guider to encode physiological responses that are acquired while humans experience VR videos. Each physiology sickness feature extractor (EEG, ECG, and GSR) in the physiological response guider is designed to suit their physiological characteristics. Extracted physiology sickness features are then fused into a deep physiology feature that comprehensively reflects individual deviations of VR sickness. Finally, the VR sickness predictor assesses individual VR sickness effectively with the fusion of the deep stimulus feature and the deep physiology feature. To validate the proposed method extensively, we built two benchmark datasets which contain 360-degree VR videos with physiological responses (EEG, ECG, and GSR) and SSQ scores. Experimental results show that the proposed method achieves meaningful correlations with human SSQ scores. Further, we validate the effectiveness of the proposed network designs by conducting analysis on feature fusion and visualization. [ABSTRACT FROM AUTHOR]

As most human actions in video sequences embody the continuous interactions between foregrounds rather than the background scene, it is significant to disentangle these foregrounds from the background for advanced action recognition systems. In this paper, therefore, we propose a Foreground EXtraction (FEX) block to explicitly model the foreground clues to achieve effective management of action subjects. In particular, the designed FEX block contains two components. The first part is a Foreground Enhancement (FE) module, which highlights the potential feature channels related to the action attributes, providing channel-level refinement for the following spatiotemporal modeling. The second phase is a Scene Segregation (SS) module, which splits feature maps into foreground and background. Specifically, a temporal model with dynamic enhancement is constructed for the foreground part, reflecting the essential nature of the action category. While the background is modeled using simple spatial convolutions, mapping the inputs to the consistent feature space. The FEX blocks can be inserted into existing 2D CNNs (denoted as FEXNet) for spatiotemporal modeling, concentrating on the foreground clues for effective action inference. Our experiments performed on Something-Something V1, V2 and Kinetics400 verify the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

In reversible data hiding (RDH) schemes, how to select those smooth pixels, pixel pairs or pixel blocks in order to improve performance is an important issue. For pixel-value-ordering (PVO) based RDH schemes, two existing techniques appear to be defective since only two reference pixels in a block or the right and bottom neighbors of a block are exploited as the context for block selection, and their performance might be only adequate for those smooth images. For rough images, the embedding performance could be affected. In this paper, an efficient block selection method by computing a block’s smoothness with a full-enclosing context (FEC) way is proposed. Obtained results show that the proposed FEC strategy can better estimate a block’s smoothness for PVO-based schemes. Furthermore, a more scalable pairing way is presented for the recently reported location-based PVO predictor. The proposed PVO scheme can be implemented by dividing the cover image and embedding bits into two different types of blocks, respectively. Experimental results show that the marked images by the proposed two-stage PVO scheme have higher visual quality, e.g., the average PSNR for the Kodak image database is 63.31 dB after embedding 10,000 bits, and the gain is 0.16 dB against the best result in the literature. Compared with some state-of-the-art RDH works, the superiority of the proposed algorithm has been verified in extensive experiments. [ABSTRACT FROM AUTHOR]

Efficient real-time disparity estimation is critical for the application of stereo vision systems in various areas. Recently, stereo network based on coarse-to-fine method has largely relieved the memory constraints and speed limitations of large-scale network models. Nevertheless, all of the previous coarse-to-fine designs employ constant offsets and three or more stages to progressively refine the coarse disparity map, still resulting in unsatisfactory computation accuracy and inference time when deployed on mobile devices. This paper claims that the coarse matching errors can be corrected efficiently with fewer stages as long as more accurate disparity candidates can be provided. Therefore, we propose a dynamic offset prediction module to meet different correction requirements of diverse objects and design an efficient two-stage framework. In addition, a disparity-independent convolution is proposed to regularize the compact cost volume efficiently and further improve the overall performance. The disparity quality and efficiency of various stereo networks are evaluated on multiple datasets and platforms. Evaluation results demonstrate that, the disparity error rate of the proposed network achieves 2.66% and 2.71% on KITTI 2012 and 2015 test sets respectively, where the computation speed is $2\times $ faster than the state-of-the-art lightweight models on high-end and source-constrained GPUs. [ABSTRACT FROM AUTHOR]

During the past few years, face videos, e.g., video conference, interviews and variety shows, have grown explosively with millions of users over social media networks. Unfortunately, the existing compression algorithms are applied to these videos for reducing bandwidth, which also bring annoying artifacts to face regions. This paper addresses the problem of face quality enhancement in compressed videos by reducing the artifacts of face regions. Specifically, we establish a compressed face video (CFV) database, which includes 196,337 faces in 214 high-quality video sequences and their corresponding 1,712 compressed sequences. We find that the faces of compressed videos exhibit tremendous scale variation and quality fluctuation. Motivated by scalable video coding, we propose a multi-scale recurrent scalable network (MRS-Net+) to enhance the quality of multi-scale faces in compressed videos. The MRS-Net+ is comprised by one base and two refined enhancement levels, corresponding to the quality enhancement of small-, medium- and large-scale faces, respectively. In the multi-level architecture of our MRS-Net+, small-/medium-scale face quality enhancement serves as the basis for facilitating the quality enhancement of medium-/large-scale faces. We further develop a landmark-assisted pyramid alignment (LPA) subnet to align faces across consecutive frames, and then apply the mask-guided quality enhancement (QE) subnet for enhancing multi-scale faces. Finally, experimental results show that our MRS-Net+ method achieves averagely 1.196 dB improvement of peak signal-to-noise ratio (PSNR) and 23.54% saving of Bjøntegaard distortion-rate (BD-rate), significantly outperforming other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

This paper presents FastInter360, a fast inter mode decision algorithm for accelerating the encoding of ERP 360 videos. The development of FastInter360 involves an in-depth and comprehensive set of evaluations performed to understand the differences in the encoder’s behavior when encoding 360 and conventional videos. These evaluations showed that due to the texture distortions resulting from projection, the encoder presents a specific behavior when encoding 360 videos, making it more likely to use a recurrent set of encoding modes when processing 360 videos. Besides, the coding efficiency is less sensible to approximations in some encoding steps depending on the frame region. FastInter360 is then proposed to reduce the encoding complexity by exploiting these differences. FastInter360 comprises three algorithms that accelerate the encoding by performing early decision by SKIP mode, reducing integer motion estimation search range, and adjusting fractional motion estimation precision. Furthermore, each of these algorithms behaves according to distortion intensity, performing greater complexity reduction in more distorted regions. When employed altogether, these algorithms compose FastInter360, which is able to achieve an average complexity reduction of 22.84% with a coding efficiency loss of 0.652% BD-BR, on average, making FastInter360 competitive with literature works. [ABSTRACT FROM AUTHOR]

The triplet loss is widely used in learning the local descriptors for image matching. However, existing triplet loss-based methods, like HardNet and DSM, employ the point-to-point distance metric, which neglects the neighborhood information of descriptors. Considering the fact that local neighborhood structures of matching descriptors would be similar under the ideal condition, this paper aims to learn the neighborhood topology-consistent descriptors (TCDesc). To this end, we first propose the linear combination weight as the topology weight to depict the neighborhood topology for each descriptor, where the difference between the center descriptor and the linear combination of its neighbors is minimized. For the global comparison, we then define a global topology vector by using the local topology weights. Next, beyond the Euclidean distance, we define a topology distance with the topology vectors to indicate the topological difference between the matching descriptors. Furthermore, we propose an adaptive weighting strategy to jointly minimize the topology distance and Euclidean distance in triplet loss. Experimental results on four widely-used datasets, i.e., UBC PhotoTourism, HPatches, W1BS and Oxford, demonstrate that our method can effectively improve the performance of both HardNet and DSM. [ABSTRACT FROM AUTHOR]

With the prevalence of accessible multi-modal sensors and the maturity of pose estimation algorithms, skeleton-based action recognition has gradually become the mainstream of human action recognition (HAR). The key issue is to mine the correlations and dependencies between different joints and bones. In this paper, we propose a cross view learning approach. First, the static and dynamic representations of skeletons, from two different views (joints and bones), are calculated and aggregated respectively. Then, the integrated representations of these two views are used as parallel inputs to the cross view learning model, which mainly includes two blocks, namely a multi-scale learning block and a multi-view fusion block. The former is used to excavate the intra-view’s discriminative and comprehensive features, and the latter is utilized to capture the complementary representations of the inter-view. Finally, the fused representations are input to the classifier for action recognition. It has been experimentally proven that our proposed approach outperforms several state-of-the-art baseline methods and achieves a very competitive performance. [ABSTRACT FROM AUTHOR]

Due to the importance of practical applications, unsupervised domain adaptation (UDA) person re-identification (re-ID) has attracted increasing attention. However, most of existing methods often lack the multi-view information reasoning and ignore the domain discrepancy of the pedestrian images with the same identity, which constrain the further improvement of recognition performance. So, this paper proposes a triple adversarial learning and multi-view imaginative reasoning network (TAL-MIRN) for UDA person re-ID, which consists of a multi-view imaginative reasoning module (IRM) and a triple adversarial learning module (TALM). IRM makes the classified pedestrian identity features from a single-view image extracted by a feature encoder consistent with the classification results of the aggregated multi-view pedestrian identity features, so the strong multi-view imaginative reasoning ability of the feature encoder is obtained. TALM is composed by the adversarial learning between the camera classifier and feature encoder, adversarial learning of joint distribution alignment, and adversarial learning of the difference between two classifiers used in classification. In particular, the domain-invariant features at camera level are guaranteed by the adversarial learning between the feature extractor and camera classifier. The joint alignment of identity and domain is achieved by the competition between the feature extractor and classifier integrated with identity and domain. The discriminability and robustness of the learned features are enhanced by playing a MinMax game between two different identity classifiers. Furthermore, a simple normalization operation named as cross normalization (CN) is proposed to increase both modeling and generalization capability of the proposed TAL-MIRN across multiple domains. The proposed TAL-MIRN is applied to five benchmark datasets, and the comparative experimental results confirm its superiority over the state-of-the-art methods. The related source codes is available at https://github.com/lhf12278/TALM-IRM. [ABSTRACT FROM AUTHOR]

The rich semantic information contained in images is an important clue to explore visual emotions. Therefore, exploring the correlation between visual emotion and the semantic relationship of objects, and extracting more effective semantic features through explicit or implicit modeling is very important for visual emotion analysis. In this paper, a novel Graph-based Object Semantic Refinement (GOSR) model is proposed to extract multi-level semantic features for visual emotion classification, in which graph structures is used to represent the object semantics and their position relationships of an image, and Graph Convolutional Networks (GCN) is used to refine object information by the aggregating neighbor object with their position relationships. The different convolutional layer’s features from GCN are further fused by Gated Recurrent Units (GRU) networks to achieve high-level semantic features. Then a framework with two branches to leverage visual and semantic information for visual sentiment analysis is proposed, which uses convolutional neural networks to extract visual features from images, and collaborates with semantic features from GOSR model to achieve better emotion recognition results. Besides, for alleviating the potentially unreasonable predictions and promote models collaboration, a novel tendency loss function based on the correlations among emotion labels is proposed to adjust the output activation value other than the target label. Extensive experiments on four widely used benchmark datasets show that our proposed method can achieve competitive performance and outperform most of the state-of-the-art methods on visual emotion recognition. [ABSTRACT FROM AUTHOR]

Unsupervised 3D model analysis has attracted tremendous attentions with the increasing growth of 3D model data and the extensive human annotations. Many effective methods have been designed to address the 3D model analysis with labeled information, while rare methods devote to unsupervised deep learning due to the difficulty of mining reliable information. In this paper, we propose a novel unsupervised deep learning method named joint local correlation and global contextual information (LCGC) for 3D model retrieval and classification, which mines the reliable triplet set and uses triplet loss to optimize the deep neural network. Our method proposes two schemes: 1) Local self-correlation information learning, which adopts the intra and inter information to construct the view-level triplet set. 2) Global neighbor contextual information learning, which employs the neighbor contextual information to explore the reliable relations among 3D models and construct the model-level triplet set. The above schemes encourage that the selected triple set can been used to improve the discrimination of learned features. Extensive evaluations on two large-scale datasets, ModelNet40 and ShapeNet55, have demonstrated the effectiveness of our proposed method. [ABSTRACT FROM AUTHOR]

Skeleton-based human action recognition has received extensive attention due to its efficiency and robustness to complex backgrounds. Though the human skeleton can accurately capture the dynamics of human poses, it fails to recognize human actions induced by the interaction between human and objects, making it is of great importance to further explore the interaction between the human and objects for human action recognition. In this paper, we devise the multi-stream interaction networks (MSIN), to simultaneously explore the dynamics of human skeleton, objects, and the interaction between human and objects. Specifically, apart from the traditional human skeleton stream, 1) the second stream explores the dynamics of object appearance from the objects surrounding the human body joints; and 2) the third stream captures the dynamics of object position in regard to the distance between the object and different human body joints. Experimental results on three popular skeleton-based human action recognition datasets, NTU RGB + D, NTU RGB + D 120, and SYSU, demonstrate the effectiveness of the proposed method, especially for recognizing the human actions with human-object interactions. [ABSTRACT FROM AUTHOR]

Previous video salient object detection (VSOD) approaches have mainly focused on the perspective of network design for achieving performance improvements. However, with the recent slowdown in the development of deep learning techniques, it might become increasingly difficult to anticipate another breakthrough solely via complex networks. Therefore, this paper proposes a universal learning scheme to obtain a further 3% performance improvement for all state-of-the-art (SOTA) VSOD models. The major highlight of our method is that we propose the ‘motion quality’, a new concept for mining video frames from the ‘buffered’ testing video stream for constructing a fine-tuning set. By using our approach, all frames in this set can all well-detect their salient object by the ‘target SOTA model’ — the one we want to improve. Thus, the VSOD results of the mined set, which were previously derived by the target SOTA model, can be directly applied as pseudolearning objectives to fine-tune a completely new spatial model that has been pretrained on the widely used DAVIS-TR set. Since some spatial scenes in the buffered testing video stream are shown, the fine-tuned spatial model can perform very well for the remaining unseen testing frames, outperforming the target SOTA model significantly. Although offline model fine tuning requires additional time costs, the performance gain can still benefit scenarios without speed requirements. Moreover, its semisupervised methodology might have considerable potential to inspire the VSOD community in the future. [ABSTRACT FROM AUTHOR]