Search

Personality determines various human daily and working behaviours. Recently, a large number of automatic personality computing approaches have been developed to predict either the apparent or self-reported personality of the subject based on non-verbal audio-visual behaviours. However, most of them suffer from complex and dataset-specific pre-processing steps and model training tricks. In the absence of a standardized benchmark with consistent experimental settings, it is not only impossible to fairly compare the real performances of these personality computing models but also makes them difficult to be reproduced. This paper presents the first reproducible audio-visual benchmark to provide a fair and consistent evaluation of eight existing personality computing models (e.g., audio, visual and audio-visual) and seven standard deep learning models on both self-reported and apparent personality recognition tasks. Building upon a set of benchmarked models, we also investigate the impact of two previously-used long-term modelling strategies for summarising short-term/frame-level predictions on personality computing results. We comprehensively investigate all benchmarked models on two publicly available datasets, ChaLearn First Impression and UDIVA self-reported personality datasets, and conclude: (i) apparent personality traits, inferred from facial behaviours by most benchmarked deep learning models, show more reliability than self-reported ones; (ii) visual models frequently achieved superior performances than audio models on personality recognition; (iii) non-verbal behaviours contribute differently in predicting different personality traits; and (iv) our reproduced personality computing models generally achieved worse performances than their original reported results. [ABSTRACT FROM AUTHOR]

Hidayetoğlu et al. propose a novel memory-centric algorithm to reconstruct X-ray CT images in the SC19 article entitled “MemXCT: Memory-Centric X-ray CT Reconstruction with Massive Parallelization”. They formulate the reconstruction with several SpMVs, and propose two memory-centric optimizations to improve cache locality for better memory bandwidth utilization, i.e., a two-level pseudo-Hilbert ordering and a multi-stage input buffering. In this article, we present our results on reproducing that article to show its effectiveness and generality, as part of the SC20 Student Cluster Competition Reproducibility Challenge. We reproduce the execution time and memory bandwidth tests in that article on various architectures, including Intel CPUs, AMD CPUs, and NVIDIA GPUs. We further analyze the bottleneck on different architectures by comparing the achieved memory bandwidth with the peak bandwidth on those architectures. We then reproduce the strong scaling test on CPU and GPU clusters with different scales, and use the proposed algorithm to reconstruct three new X-ray computed tomograms. [ABSTRACT FROM AUTHOR]

Meta-heuristic algorithms, especially evolutionary algorithms, have been frequently used to find near optimal solutions to combinatorial optimization problems. The evaluation of such algorithms is often conducted through comparisons with other algorithms on a set of benchmark problems. However, even if one algorithm is the best among all those compared, it still has difficulties in determining the true quality of the solutions found because the true optima are unknown, especially in dynamic environments. It would be desirable to evaluate algorithms not only relatively through comparisons with others, but also in absolute terms by estimating their quality compared to the true global optima. Unfortunately, true global optima are normally unknown or hard to find since the problems being addressed are NP-hard. In this paper, instead of using true global optima, lower bounds are derived to carry out an objective evaluation of the solution quality. In particular, the first approach capable of deriving a lower bound for dynamic capacitated arc routing problem (DCARP) instances is proposed, and two optimization algorithms for DCARP are evaluated based on such a lower bound approach. An effective graph pruning strategy is introduced to reduce the time complexity of our proposed approach. Our experiments demonstrate that our approach provides tight lower bounds for small DCARP instances. Two optimization algorithms are evaluated on a set of DCARP instances through the derived lower bounds in our experimental studies, and the results reveal that the algorithms still have room for improvement for large complex instances. [ABSTRACT FROM AUTHOR]

This paper introduces video domain generalization where most video classification networks degenerate due to the lack of exposure to the target domains of divergent distributions. We observe that the global temporal features are less generalizable, due to the temporal domain shift that videos from other unseen domains may have an unexpected absence or misalignment of the temporal relations. This finding has motivated us to solve video domain generalization by effectively learning the local-relation features of different timescales that are more generalizable, and exploiting them along with the global-relation features to maintain the discriminability. This paper presents the VideoDG framework with two technical contributions. The first is a new deep architecture named the Adversarial Pyramid Network, which improves the generalizability of video features by capturing the local-relation, global-relation, and cross-relation features progressively. On the basis of pyramid features, the second contribution is a new and robust approach of adversarial data augmentation that can bridge different video domains by improving the diversity and quality of augmented data. We construct three video domain generalization benchmarks in which domains are divided according to different datasets, different consequences of actions, or different camera views, respectively. VideoDG consistently outperforms the combinations of previous video classification models and existing domain generalization methods on all benchmarks. [ABSTRACT FROM AUTHOR]

The vanilla Few-shot Learning (FSL) learns to build a classifier for a new concept from one or very few target examples, with the general assumption that source and target classes are sampled from the same domain. Recently, the task of Cross-Domain Few-Shot Learning (CD-FSL) aims at tackling the FSL where there is a huge domain shift between the source and target datasets. Extensive efforts on CD-FSL have been made via either directly extending the meta-learning paradigm of vanilla FSL methods, or employing massive unlabeled target data to help learn models. In this paper, we notice that in the CD-FSL task, the few labeled target images have never been explicitly leveraged to inform the model in the training stage. However, such a labeled target example set is very important to bridge the huge domain gap. Critically, this paper advocates a more practical training scenario for CD-FSL. And our key insight is to utilize a few labeled target data to guide the learning of the CD-FSL model. Technically, we propose a novel Generalized Meta-learning based Feature-Disentangled Mixup network, namely GMeta-FDMixup. We make three key contributions of utilizing GMeta-FDMixup to address CD-FSL. Firstly, we present two mixup modules – mixup-P and mixup-M that help facilitate utilizing the unbalanced and disjoint source and target datasets. These two novel modules enable diverse image generation for training the model on the source domain. Secondly, to narrow the domain gap explicitly, we contribute a novel feature disentanglement module that learns to decouple the domain-irrelevant and domain-specific features. By stripping the domain-specific features, we alleviate the negative effects caused by the domain inductive bias. Finally, we repurpose a new contrastive learning module, dubbed ConL. ConL prevents the model from only capturing category-related features via introducing contrastive loss. Thus, the generalization ability on novel categories is improved. Extensive experimental results on two benchmarks show the superiority of our setting and the effectiveness of our method. Code and models will be released. [ABSTRACT FROM AUTHOR]

Extracting robust and discriminative local features from images plays a vital role for long term visual localization, whose challenges are mainly caused by the severe appearance differences between matching images due to the day-night illuminations, seasonal changes, and human activities. Existing solutions resort to jointly learning both keypoints and their descriptors in an end-to-end manner, leveraged on large number of annotations of point correspondence which are harvested from the structure from motion and depth estimation algorithms. While these methods show improved performance over non-deep methods or those two-stage deep methods, i.e., detection and then description, they are still struggled to conquer the problems encountered in long term visual localization. Since the intrinsic semantics are invariant to the local appearance changes, this paper proposes to learn semantic-aware local features in order to improve robustness of local feature matching for long term localization. Based on a state of the art CNN architecture for local feature learning, i.e., ASLFeat, this paper leverages on the semantic information from an off-the-shelf semantic segmentation network to learn semantic-aware feature maps. The learned correspondence-aware feature descriptors and semantic features are then merged to form the final feature descriptors, for which the improved feature matching ability has been observed in experiments. In addition, the learned semantics embedded in the features can be further used to filter out noisy keypoints, leading to additional accuracy improvement and faster matching speed. Experiments on two popular long term visual localization benchmarks (Aachen Day and Night v1.1, Robotcar Seasons) and one challenging indoor benchmark (InLoc) demonstrate encouraging improvements of the localization accuracy over its counterpart and other competitive methods. [ABSTRACT FROM AUTHOR]

Cooperative co-evolution (CC) is an evolutionary algorithm that adopts the divide-and-conquer strategy to solve large-scale optimization problems. It is difficult for CC to specify a suitable subpopulation size to solve different subproblems. The population diversity may be insufficient to search for the global optimum during subpopulations' evolution. In this paper, an adaptive method for enhancing population diversity is embedded in a contribution-based CC. In CC, there are two kinds of subpopulation: the convergent or stagnant subpopulations and the non-convergent and non-stagnant subpopulations. A method is proposed in the paper to evaluate the convergent and stagnant subpopulations' contributions to improving the best overall objective value, which is different from the contribution evaluation on the non-convergent and non-stagnant subpopulations. In each co-evolutionary cycle, the new CC adaptively determines to select a subpopulation, which can make a greater contribution to improving the best overall objective value, between the above two kinds of subpopulation to undergo evolution. When a convergent or stagnant subpopulation is selected to undergo evolution, the subpopulation is re-diversified to enhance its global search capability. Our experimental results and analysis suggest that the new CC algorithm can improve the performance of CC and serves as a competitive solver for large-scale optimization problems. [ABSTRACT FROM AUTHOR]

RDF is being increasingly considered in a broad range of information management scenarios. Governments, large corporations, startups, and other organizations around the world are using RDF as a data model to represent and share knowledge. However, there is still a long evolutionary track with multiple challenges for RDF reaching the scale of the most recent Big Data intensive applications (e.g., Smart Cities, Sensor Networks, eHealth, Internet of Things). In this survey, we review the usage of NoSQL databases to the storage of large RDF graphs by rehearsing the latest surveys and expanding their findings by updating proposals and bringing light to aspects such as model mapping between RDF and NoSQL, triple indexing and partitioning, graph fragmentation and data caching. Moreover, we explain how the surveyed works extended the RDF capabilities so the datasets can benefit of the characteristics of scalability, schemaless data, and better overall performance of NoSQL databases. The survey summarizes the current state of art, discusses open problems, and proposes a Reference Architecture (RA). For the best of our knowledge, this is the first survey where the focus is solely on papers that use one or more NoSQL systems for the RDF persistence. [ABSTRACT FROM AUTHOR]

In modern industries, data-driven fault detection and classification (FDC) systems can efficiently maintain industrial security and stability, while the security of the data-driven FDC system itself is rarely or even never considered. The security problem named adversarial vulnerability is the intrinsic of data-driven machine learning models, which will give incorrect predictions under the maliciously perturbed input data. This paper presents a work on this new security topic of the data-driven FDC systems, by 1) summarizing and comparing various recent and typical adversarial attack and defense methods for fault classifiers; 2) proposing novel attack and defense techniques for unsupervised fault detectors; 3) constructing a novel industrial adversarial security benchmark on FDC systems in the Tennessee-Eastman process (TEP) dataset; 4) exploring and discussing which attack is most potentially threatening for FDC systems and which defense technique is most applicable to mitigate attacks. The results reveal unique security properties of FDC systems, mainly including 1) for fault classifiers, black-box attack is close to the attack strength of white-box FGSM and the universal transferable attack is not significantly stronger than random noise; 2) weak adversarial training is excellent with high adversarial accuracy improvement and negligible clean accuracy decrease; 3) fault detectors are intrinsically more robust, and can be well protected by strong adversarial training. More intriguing properties and profound insights are demonstrated in the paper. This pioneering work could guide researchers and practitioners in discovering and navigating the field of FDC system adversarial robustness, outlining the research directions and open problems. [ABSTRACT FROM AUTHOR]

Recent work has shown that convolutional networks can be substantially deeper, more accurate, and efficient to train if they contain shorter connections between layers close to the input and those close to the output. In this paper, we embrace this observation and introduce the Dense Convolutional Network (DenseNet), which connects each layer to every other layer in a feed-forward fashion. Whereas traditional convolutional networks with $L$ L layers have $L$ L connections—one between each layer and its subsequent layer—our network has $\frac{L(L+1)}{2}$ L (L + 1) 2 direct connections. For each layer, the feature-maps of all preceding layers are used as inputs, and its own feature-maps are used as inputs into all subsequent layers. DenseNets have several compelling advantages: they alleviate the vanishing-gradient problem, encourage feature reuse and substantially improve parameter efficiency. We evaluate our proposed architecture on four highly competitive object recognition benchmark tasks (CIFAR-10, CIFAR-100, SVHN, and ImageNet). DenseNets obtain significant improvements over the state-of-the-art on most of them, whilst requiring less parameters and computation to achieve high performance. [ABSTRACT FROM AUTHOR]

Weight sharing promises to make neural architecture search (NAS) tractable even on commodity hardware. Existing methods in this space rely on a diverse set of heuristics to design and train the shared-weight backbone network, a.k.a. the super-net. Since heuristics substantially vary across different methods and have not been carefully studied, it is unclear to which extent they impact super-net training and hence the weight-sharing NAS algorithms. In this paper, we disentangle super-net training from the search algorithm, isolate 14 frequently-used training heuristics, and evaluate them over three benchmark search spaces. Our analysis uncovers that several commonly-used heuristics negatively impact the correlation between super-net and stand-alone performance, whereas simple, but often overlooked factors, such as proper hyper-parameter settings, are key to achieve strong performance. Equipped with this knowledge, we show that simple random search achieves competitive performance to complex state-of-the-art NAS algorithms when the super-net is properly trained. [ABSTRACT FROM AUTHOR]

Timing-based side and covert channels in processor caches continue to be a threat to modern computers. This work shows for the first time, a systematic, large-scale analysis of Arm devices and the detailed results of attacks the processors are vulnerable to. Compared to x86, Arm uses different architectures, microarchitectural implementations, cache replacement policies, etc., which affects how attacks can be launched, and how security testing for the vulnerabilities should be done. To evaluate security, this paper presents security benchmarks specifically developed for testing Arm processors and their caches. The benchmarks are evaluated with sensitivity tests, which examine how sensitive the benchmarks are to having a correct configuration in the testing phase. Further, to evaluate a large number of devices, this work leverages a novel approach of using a cloud-based Arm device testbed for architectural and security research on timing channels and runs the benchmarks on 34 different physical devices. In parallel, there has been much interest in secure caches to defend the various attacks. Consequently, this paper also investigates secure cache architectures using proposed benchmarks. Especially, this paper implements and evaluates secure PL and RF caches, showing the security of PL and RF caches, but also uncovers new weaknesses. [ABSTRACT FROM AUTHOR]

Previous studies have underlined the importance of establishing a database of high‐fidelity benchmark experimental test results under random loading histories such as earthquake excitations for both existing structural systems as well as newly proposed ones. Such experimental results are useful for understanding the limitations of previously developed numerical models, proposing new numerical models for more recently developed structural systems, better assessing the ultra‐low cycle fatigue (ULCF) life of energy dissipative components, and improving the loading protocols used for element assessments or developments of numerical model. This paper presents experimental results from pseudo‐dynamic hybrid simulations on steel eccentrically braced frames (EBF) equipped with novel cast steel replaceable modular yielding links. The first set of experiments are carried out on a four‐story EBF, with the first‐floor yielding link physically tested in the laboratory. The second set of experiments are performed on a two‐story EBF where the second‐floor link is physically tested in the laboratory, while considering axial loads due to the imbalanced distribution of the seismic weight in the building. Each building structure is subjected to three Maximum Credible Event (MCE) level earthquakes. A framework for performing hybrid simulations on EBFs is proposed where the response of the yielding link is physically captured in a single degree of freedom control system, with or without axial loads. A substructuring strategy is proposed in conjunction with a modeling approach in the integration module. The results confirm the effectiveness of the hybrid simulation framework, which can be adopted in similar studies. Two numerical models are proposed for capturing the response of cast steel links using a simplified and a more advanced approach. The models are first calibrated using incremental reversed‐cyclic experimental test results, and then refined using the hybrid simulation test results. The effects of this calibration improvement are quantified through critically comparing the response prediction for different response parameters before and after improving the calibration with the experimental results. The numerical model which captures the physical mechanism of the cast steel links provides a more accurate response prediction compared to the simplified numerical model using a phenomenological truss with calibrated force‐deformation. The remaining ULCF life of the tested yielding links is experimentally investigated after having sustained three MCE level earthquakes, which indicates a large reserve ductility capacity which is more than twice the required plastic rotation capacity for cast steel links after seismic events. [ABSTRACT FROM AUTHOR]

In this paper, we make the first benchmark effort to elaborate on the superiority of using RAW images in the low light enhancement and develop a novel alternative route to utilize RAW images in a more flexible and practical way. Inspired by a full consideration on the typical image processing pipeline, we are inspired to develop a new evaluation framework, Factorized Enhancement Model (FEM), which decomposes the properties of RAW images into measurable factors and provides a tool for exploring how properties of RAW images affect the enhancement performance empirically. The empirical benchmark results show that the Linearity of data and Exposure Time recorded in meta-data play the most critical role, which brings distinct performance gains in various measures over the approaches taking the sRGB images as input. With the insights obtained from the benchmark results in mind, a RAW-guiding Exposure Enhancement Network (REENet) is developed, which makes trade-offs between the advantages and inaccessibility of RAW images in real applications in a way of using RAW images only in the training phase. REENet projects sRGB images into linear RAW domains to apply constraints with corresponding RAW images to reduce the difficulty of modeling training. After that, in the testing phase, our REENet does not rely on RAW images. Experimental results demonstrate not only the superiority of REENet to state-of-the-art sRGB-based methods and but also the effectiveness of the RAW guidance and all components. [ABSTRACT FROM AUTHOR]

Single-guide RNA is a guide RNA (gRNA), which guides the insertion or deletion of uridine residues into kinetoplastid during RNA editing. It is a small non-coding RNA that can be combined with pre -mRNA pairing. SgRNA is a critical component of the CRISPR/Cas9 gene knockout system and play an important role in gene editing and gene regulation. It is important to accurately and quickly identify highly on-target activity sgRNAs. Due to its importance, several computational predictors have been proposed to predict sgRNAs on-target activity. All these methods have clearly contributed to the development of this very important field. However, they also have certain limitations. In the paper, we developed a new classifier SgRNA-RF, which extracts the features of nucleic acid composition and structure of on-target activity sgRNA sequence and identified by random forest algorithm. In addition to solving an imbalanced dataset, this paper proposed a new method called CS-Smote. We compared sgRNA-RF with state-of-the-art predictors on the five datasets, and found SgRNA-RF significantly improved the identification accuracy, with accuracies of 0.8636,0.9161,0.894,0.938,0.965,0.77,0.979,0.973, respectively. The user-friendly web server that implements sgRNA-RF is freely available at http://server.malab.cn/sgRNA-RF/. [ABSTRACT FROM AUTHOR]

Full projector compensation aims to modify a projector input image to compensate for both geometric and photometric disturbance of the projection surface. Traditional methods usually solve the two parts separately and may suffer from suboptimal solutions. In this paper, we propose the first end-to-end differentiable solution, named CompenNeSt++, to solve the two problems jointly. First, we propose a novel geometric correction subnet, named WarpingNet, which is designed with a cascaded coarse-to-fine structure to learn the sampling grid directly from sampling images. Second, we propose a novel photometric compensation subnet, named CompenNeSt, which is designed with a siamese architecture to capture the photometric interactions between the projection surface and the projected images, and to use such information to compensate the geometrically corrected images. By concatenating WarpingNet with CompenNeSt, CompenNeSt++ accomplishes full projector compensation and is end-to-end trainable. Third, to improve practicability, we propose a novel synthetic data-based pre-training strategy to significantly reduce the number of training images and training time. Moreover, we construct the first setup-independent full compensation benchmark to facilitate future studies. In thorough experiments, our method shows clear advantages over prior art with promising compensation quality and meanwhile being practically convenient. [ABSTRACT FROM AUTHOR]

Visual localization enables autonomous vehicles to navigate in their surroundings and augmented reality applications to link virtual to real worlds. Practical visual localization approaches need to be robust to a wide variety of viewing conditions, including day-night changes, as well as weather and seasonal variations, while providing highly accurate six degree-of-freedom (6DOF) camera pose estimates. In this paper, we extend three publicly available datasets containing images captured under a wide variety of viewing conditions, but lacking camera pose information, with ground truth pose information, making evaluation of the impact of various factors on 6DOF camera pose estimation accuracy possible. We also discuss the performance of state-of-the-art localization approaches on these datasets. Additionally, we release around half of the poses for all conditions, and keep the remaining half private as a test set, in the hopes that this will stimulate research on long-term visual localization, learned local image features, and related research areas. Our datasets are available at visuallocalization.net , where we are also hosting a benchmarking server for automatic evaluation of results on the test set. The presented state-of-the-art results are to a large degree based on submissions to our server. [ABSTRACT FROM AUTHOR]

Previous studies have demonstrated that the response of stable yielding systems to random excitations such as earthquakes is inherently accompanied by residual deformations. Stable yielding systems with highly enhanced ductility capacity are more prone to residual deformations. Steel eccentrically braced frames (EBFs) are designed to sustain significant localized inelastic deformations in the segment of the beam referred to as the yielding link, which results in much larger link rotations when compared to inter‐story drift ratios (SDR). Recent experimental and numerical studies on the response of EBFs have demonstrated that even with moderate residual inter‐SDR (less than 0.5%), severe link residual rotations could be expected, which could render the structure difficult to repair. For replaceable yielding links, which were developed to expedite the repair and recovery process in EBFs after major seismic events, large residual link rotations will hinder the link replacement process or at times make it impossible. Therefore, mitigation of residual deformations is of greater importance in EBFs with replaceable yielding links than in conventional EBFs with a continuous beam. This paper provides a brief review of past research on residual deformations and methods for their mitigation. A low‐cost re‐centering system for use in EBFs is proposed to work in parallel with the yielding link, in order to mitigate residual deformations. The proposed system is adopted for two prototype structures designed with EBFs equipped with cast steel replaceable modular yielding links. The effectiveness of the re‐centering strategy is evaluated by means of pseudo‐dynamic hybrid simulations, as well as numerical simulations. The hybrid simulations and numerical studies both demonstrate that not only the proposed system is highly effective at mitigating the residual deformations, but also it is effective in reducing the peak deformations in EBFs. [ABSTRACT FROM AUTHOR]

Unsupervised domain adaptation addresses the problem of transferring knowledge from a well-labeled source domain to an unlabeled target domain where the two domains have distinctive data distributions. Thus, the essence of domain adaptation is to mitigate the distribution divergence between the two domains. The state-of-the-art methods practice this very idea by either conducting adversarial training or minimizing a metric which defines the distribution gaps. In this paper, we propose a new domain adaptation method named adversarial tight match (ATM) which enjoys the benefits of both adversarial training and metric learning. Specifically, at first, we propose a novel distance loss, named maximum density divergence (MDD), to quantify the distribution divergence. MDD minimizes the inter-domain divergence (“match” in ATM) and maximizes the intra-class density (“tight” in ATM). Then, to address the equilibrium challenge issue in adversarial domain adaptation, we consider leveraging the proposed MDD into adversarial domain adaptation framework. At last, we tailor the proposed MDD as a practical learning loss and report our ATM. Both empirical evaluation and theoretical analysis are reported to verify the effectiveness of the proposed method. The experimental results on four benchmarks, both classical and large-scale, show that our method is able to achieve new state-of-the-art performance on most evaluations. [ABSTRACT FROM AUTHOR]

Several optimization problems have features that hinder the capabilities of searching heuristics. To cope with this issue, different methods have been proposed to manipulate search spaces and improve the optimization process. This paper focuses on Dilation Functions (DFs), which are one of the most promising techniques to manipulate the fitness landscape, by "expanding" or "compressing" specific regions. The definition of appropriate DFs is problem dependent and requires a-priori knowledge of the optimization problem. Therefore, it is essential to introduce an automatic and efficient strategy to identify optimal DFs. With this aim, we propose a novel method based on Genetic Programming, named GP4DFs, which is capable of evolving effective DFs. GP4DFs identifies optimal dilations, where a specific DF is applied to each dimension of the search space. Moreover, thanks to a knowledge-driven initialization strategy, GP4DFs converges to better solutions with a reduced number of fitness evaluations, compared to the state-of-the-art approaches. The performance of GP4DFs is assessed on a set of 43 benchmark functions mimicking several features of real-world optimization problems. The obtained results indicate the suitability of the generated DFs. [ABSTRACT FROM AUTHOR]

Zero-Shot Learning (ZSL) aims to recognize unseen classes that never appear during training. Recently, generative adversarial networks (GANs) have been introduced to convert ZSL into a supervised learning problem by synthesizing unseen visual features. However, since unseen classes are never experienced for the generator during training, the synthesized unseen visual features often become heavily biased towards seen classes, or sometimes there is even no meaningful class that can be assigned to them. This is known as the bias problem. In this paper, we propose a novel method, namely Adaptive Bias-Aware GAN (ABA-GAN), to alleviate generating biased visual features. For this purpose, we build a semantic adversarial network to regularize the feature generator. Specifically, an adaptive adversarial loss is proposed to constrain the feature distributions, which avoids the generation of meaningless visual features. Meanwhile, a domain divider is presented to explicitly distinguish synthesized visual features between seen and unseen domains, such that the bias towards seen classes can be alleviated. Moreover, we propose a novel metric named bias score (BS) to explicitly quantify the degree of the strong bias. Extensive experiments on four widely used benchmark datasets demonstrate that our proposed method outperforms the state-of-the-art approaches under both ZSL and GZSL protocols. [ABSTRACT FROM AUTHOR]

Weakly supervised object detection (WSOD) aims to train object detectors by using only image-level annotations. Many recent works on WSOD adopt multiple instance detection networks (MIDN), which usually generate a certain number of proposals and regard proposal classification as a latent model learning within image classification. However, these methods tend to detect salient object, salient object parts and clustered objects due to lack of instance-level annotations during training. Thus a core issue is how to guarantee that the network learn as many objects with precise bounding boxes as possible. In this paper, we address this issue by exploiting the potential of proposal scores during training. We propose an adaptive instance refinement (AIR) framework with three novel designs, which can be integrated with MIDN into a single network. Specifically, adaptive instance mining attempts to discover all positive instances according to the score distribution of proposals and their spatial similarity. Adaptive score modulation dynamically adjusts proposal scores to make the network focus more on instances with different difficulties in different training iterations. Adaptive knowledge refinement distills important information from all previous stages by the weighted average of proposal scores. The experimental results on the PASCAL VOC 2007 and 2012 benchmarks and the MS COCO benchmark demonstrate that AIR significantly improves the performance of the original MIDN and achieves the state-of-the-art results. [ABSTRACT FROM AUTHOR]

The heterogeneity in recently published knowledge graph embedding models’ implementations, training, and evaluation has made fair and thorough comparisons difficult. To assess the reproducibility of previously published results, we re-implemented and evaluated 21 models in the PyKEEN software package. In this paper, we outline which results could be reproduced with their reported hyper-parameters, which could only be reproduced with alternate hyper-parameters, and which could not be reproduced at all, as well as provide insight as to why this might be the case. We then performed a large-scale benchmarking on four datasets with several thousands of experiments and 24,804 GPU hours of computation time. We present insights gained as to best practices, best configurations for each model, and where improvements could be made over previously published best configurations. Our results highlight that the combination of model architecture, training approach, loss function, and the explicit modeling of inverse relations is crucial for a model’s performance and is not only determined by its architecture. We provide evidence that several architectures can obtain results competitive to the state of the art when configured carefully. We have made all code, experimental configurations, results, and analyses available at https://github.com/pykeen/pykeen and https://github.com/pykeen/benchmarking. [ABSTRACT FROM AUTHOR]

The aim of this research paper is to conduct a comprehensive analysis of Rust and C, with the objective of assessing whether Rust can finally replace C as a preferred programming language. The analysis encompasses various aspects such as performance benchmarks, memory management, error handling mechanisms, security vulnerabilities, language syntax and readability, type safety and memory management, tooling and libraries support, development workflows, learning curve and community support, the Rust ecosystem, adoption in various industries, case studies of successful Rust projects, and a comparison of industry trends and language preferences. The findings and analysis highlight the strengths and considerations of each language, providing insights into the potential of Rust to replace C in different domains. [ABSTRACT FROM AUTHOR]

Multi-object tracking (MOT), which is crucial for computer vision and video processing, has immense potential for improvement. Traditional tracking-by-detection approaches include feature-based object re-identification methods that use trained features, but these methods suffer from a lack of suitable training data. In training datasets used for MOT, every object in a video sequence must have its own location and ID. However, assigning IDs to each object in every sequence is considerably labor-intensive, and hence current MOT datasets are unsuitable for training re-identification networks. To resolve this issue, this paper proposes a novel self-supervised learning method using several short videos that contain no human-added labels, based on the idea that each video is a set of temporally corresponding image frames. We then describe how to improve tracking performance using a re-identification network trained in a self-supervised manner. In addition, ablation studies were conducted in order to define the optimal parameters, such as number of clips, data augmentation, and appropriate matching algorithms. The proposed approach achieved competitive performance compared with current best-practice methods including supervised methods, achieving MOT accuracy = 62.0% and ID F1-score = 62.7% on the MOT17 benchmark. [ABSTRACT FROM AUTHOR]

Recently, angular margin-based methods have become the mainstream approach for unconstrained face recognition with remarkable success. However, robust face recognition still remains a challenge, as the face is subject to variations in pose, age, expression, occlusion, and illumination, especially in unconstrained scenarios. Since the training dataset are always collected in unconstrained scenarios, it is inevitable that there’re significant number of hard examples in the training process. In this paper, we design a hard example selection function to effectively identify hard examples in the training procedure with the supervision of angular margin-based losses. Furthermore, a novel Intra-class Hard Example Mining (IHEM) loss function is proposed, which penalizes the cosine distance between the hard examples and their class centers to enhance the discriminative power of face representations. To ensure high performance for face recognition, we combine the supervision of angular margin-based loss and IHEM loss for model training. Specifically, during the training procedure, the angular margin-based loss guarantees the power of feature discrimination for face recognition, while the IHEM loss further encourages the intra-class compactness of hard example. Extensive results demonstrate the superiority of our approach. [ABSTRACT FROM AUTHOR]

Drones, or general UAVs, equipped with cameras have been fast deployed with a wide range of applications, including agriculture, aerial photography, and surveillance. Consequently, automatic understanding of visual data collected from drones becomes highly demanding, bringing computer vision and drones more and more closely. To promote and track the developments of object detection and tracking algorithms, we have organized three challenge workshops in conjunction with ECCV 2018, ICCV 2019 and ECCV 2020, attracting more than 100 teams around the world. We provide a large-scale drone captured dataset, VisDrone, which includes four tracks, i.e., (1) image object detection, (2) video object detection, (3) single object tracking, and (4) multi-object tracking. In this paper, we first present a thorough review of object detection and tracking datasets and benchmarks, and discuss the challenges of collecting large-scale drone-based object detection and tracking datasets with fully manual annotations. After that, we describe our VisDrone dataset, which is captured over various urban/suburban areas of 14 different cities across China from North to South. Being the largest such dataset ever published, VisDrone enables extensive evaluation and investigation of visual analysis algorithms for the drone platform. We provide a detailed analysis of the current state of the field of large-scale object detection and tracking on drones, and conclude the challenge as well as propose future directions. We expect the benchmark largely boost the research and development in video analysis on drone platforms. All the datasets and experimental results can be downloaded from https://github.com/VisDrone/VisDrone-Dataset. [ABSTRACT FROM AUTHOR]

Resilient fuzzy stabilization is capable of providing much less conservative results than conventional fuzzy stabilization while the alert threshold condition should be always satisfied at each sampling instant. In order to make the alert threshold condition more easily to be guaranteed, the short paper employs the switching-type gain-scheduling control law so that the real-time information, which is specific to the current sampling instant, can be integrated into resilient fuzzy stabilization. More importantly, a new kind of time-variant balanced matrix is introduced for the first time for adjusting positive/negative terms of different monomials in a more flexible way. As a result, the conservatism of resilient fuzzy stabilization can be further reduced even if the alert threshold condition becomes more difficult to be violated. Finally, the advantage of the developed method is tested and validated via related comparisons on a benchmark example. [ABSTRACT FROM AUTHOR]

In this paper, a joint design of sparse code multiple access (SCMA) codebook and peak-to-average power ratio (PAPR) reduction based on the partial transmit sequence (PTS) technique is proposed for the downlink orthogonal frequency division multiplexing (OFDM) scheme. A benchmark scheme is firstly devised based on a direct extension of the conventional PTS-based OFDM scheme. Based on the characteristics of the near-optimal SCMA codebook designs reported in our previous work, a PTS-based SCMA-OFDM design is proposed by carefully determining the rotation angles of the SCMA codebooks, according to the adopted signal sub-constellations. A receiver which does not need the side information is further devised. Considering the phase ambiguity issue of the receiver, the PTS angles have been carefully refined. When compared to both the original scheme that does not consider PAPR reduction techniques and the benchmark scheme, the proposed scheme achieves a significantly enhanced PAPR performance with approximately the same error performance. In addition, a comparable PAPR performance, an improved error performance, and a lower computational complexity can be achieved when compared to the clipping-based SCMA-OFDM scheme reported in the previous literature. [ABSTRACT FROM AUTHOR]

Utility-scale and small scale wind and solar power installations along with electric vehicle charging stations, and other active sources of energy are increasing at the medium and lower voltage levels in the distribution grid. This situation requires a better understanding of the impact of high penetration of weather-dependent renewable energy sources on the operating conditions of the distribution network at both medium and low voltage levels. Despite the need, a multi-voltage level distribution network model, based on real network data and weather-dependent renewable generation data, has not been presented for distribution grid studies. This paper presents a comprehensive multi-voltage level active distribution network model based on real network data along with load and generation time-series for about a year. The network topology is modelled based on geographical data for various rural, semi-urban, and urban locations. The distribution network is embodied with a large share of renewable generation sources, with generation time-series simulated from meteorological data. The network is also flexible to incorporate other assets such as electric vehicle charging stations, storage, etc. The presented active distribution network model can be used to study, optimize, and control the effects of weather-dependent generation and other network assets in the distribution grid. [ABSTRACT FROM AUTHOR]

Timely updated road networks are the basis for many real-world applications such as intelligent navigation and traffic management. Existing road updating methods based on remote sensing images learn from historical road databases to update roads. Road extraction models learned from historical images however, are not easily applied to a current image due to spectral differences; and only changed roads need updating. In this paper, an Unchanged-Guided Road Updating (UGRoadUpd) framework is proposed to improve the quality of updated road networks by limiting the road updating range and learning from historical unchanged roads. The UGRoadUpd framework identifies road changes using a novel dual-task dominant-transformer-based neural network for road change detection (DT-RoadCDNet). DT-RoadCDNet executes road segmentation and change detection simultaneously, from bi-temporal remote sensing images. The Dominant-Transformer based Global Context Modeling module in DT-RoadCDNet globally models the contextual spatial structure for improved integrity in roads and road changes. Based on the discovery of road changes, an unchanged-guided road update strategy updates the roads in changed areas by learning from the prior information provided by unchanged roads in a historical road database. Experiments on two newly annotated road change detection and update datasets confirms the effectiveness of our UGRoadUpd framework. [ABSTRACT FROM AUTHOR]

Change detection is the foundation of intelligent video surveillance of the airport ground. However, experiments have shown that change detection algorithms with good performance on traditional datasets (e.g., CDnet2014) perform poorly in airport ground surveillance. The reason is that traditional datasets focus on the diversity of scenarios, while the practical application requires robustness against various changes in a single scene. We posit that the solution to this problem is to establish a unique dataset for airport ground surveillance and develop specific algorithms for this scenario. In this paper, we present an Airport Ground Video Surveillance benchmark (AGVS) for change detection of the airport ground. AGVS includes 25 long videos, amounting to about 100000 frames and accurate ground truth for all frames. Each video contains multiple challenges specific to the airport ground (e.g., haze, camouflage, strip shape, shadow and illumination change, simultaneous multi-scale objects) and various appearance changes of the aircraft). Change detection ground truth is generated by manual annotation. The AGVS benchmark can be downloaded from https://www.agvs-caac.com. Furthermore, we conduct a simple review of current change detection algorithms, both unsupervised or supervised, and then 21 state-of-the-art algorithms are tested and analyzed on the AGVS benchmark. Finally, we conclude with algorithm design principles of change detection for airport ground surveillance. [ABSTRACT FROM AUTHOR]

This paper proposes a novel numerical inverse kinematics algorithm called the Quick Inverse Kinematics or QuIK method. The QuIK method is a third-order algorithm that uses both the first- and second-order derivative information to iteratively converge to a solution. Numerical inverse kinematics methods are readily implemented on any serial robot and do not rely on joint alignment. However, they typically are slower and less robust. The second-order derivative term allows the QuIK algorithm to converge more rapidly and more robustly than existing algorithms. A damped extension to the QuIK method is also proposed to increase reliability near singularities. The QuIK methods are tested in terms of evaluation speed, reliability, and singularity robustness against the Newton–Raphson method and several other modern algorithms. The proposed QuIK methods outperform all other tested algorithms in terms of speed and robustness, and have strong performance near singularities. The QuIK algorithms are proposed as faster and more robust “drop-in” replacements to the Newton–Raphson methods in inverse kinematics. C++ and MATLAB codebases are made available. [ABSTRACT FROM AUTHOR]

With the unprecedented developments in deep learning, automatic segmentation of main abdominal organs seems to be a solved problem as state-of-the-art (SOTA) methods have achieved comparable results with inter-rater variability on many benchmark datasets. However, most of the existing abdominal datasets only contain single-center, single-phase, single-vendor, or single-disease cases, and it is unclear whether the excellent performance can generalize on diverse datasets. This paper presents a large and diverse abdominal CT organ segmentation dataset, termed AbdomenCT-1K, with more than 1000 (1K) CT scans from 12 medical centers, including multi-phase, multi-vendor, and multi-disease cases. Furthermore, we conduct a large-scale study for liver, kidney, spleen, and pancreas segmentation and reveal the unsolved segmentation problems of the SOTA methods, such as the limited generalization ability on distinct medical centers, phases, and unseen diseases. To advance the unsolved problems, we further build four organ segmentation benchmarks for fully supervised, semi-supervised, weakly supervised, and continual learning, which are currently challenging and active research topics. Accordingly, we develop a simple and effective method for each benchmark, which can be used as out-of-the-box methods and strong baselines. We believe the AbdomenCT-1K dataset will promote future in-depth research towards clinical applicable abdominal organ segmentation methods. [ABSTRACT FROM AUTHOR]

Vehicle Re-Identification (ReID) is of great significance for public security and intelligent transportation. Large and comprehensive datasets are crucial for the development of vehicle ReID in model training and evaluation. However, existing datasets in this field have limitations in many aspects, including the constrained capture conditions, limited variation of vehicle appearances, and small scale of training and test set, etc. Hence, a new, large, and challenging benchmark for vehicle ReID is urgently needed. In this paper, we propose a large vehicle ReID dataset, called VERI-Wild 2.0, containing 825,042 images. It is captured using a city-scale surveillance camera system, consisting of 274 cameras covering a very large area over 200 $km^2$ k m 2 . Specifically, the samples in our dataset present very rich appearance diversities thanks to the long time span collecting settings, unconstrained capturing viewpoints, various illumination conditions, diversified background environments, and different weather conditions. Furthermore, to facilitate more practical benchmarking, we define a challenging and large test set containing about 400K vehicle images that do not have any camera overlap with the training set. VERI-Wild 2.0 is expected to be able to facilitate the design, adaptation, development, and evaluation of different types of learning models for vehicle ReID. Besides, we also design a new method for vehicle ReID. We observe that orientation is a crucial factor for feature matching in vehicle ReID. To match vehicle pairs captured from similar orientations, the learned features are expected to capture specific detailed differential information for discriminating the visually similar yet different vehicles. In contrast, features are desired to capture the orientation invariant common information when matching samples captured from different orientations. Thus a novel disentangled feature learning network (DFNet) is proposed. It explicitly considers the orientation information for vehicle ReID, and concurrently learns the orientation specific and orientation common features that thus can be adaptively exploited via an adaptive matching scheme when dealing with matching pairs from similar or different orientations. The comprehensive experimental results show the effectiveness of our proposed method. [ABSTRACT FROM AUTHOR]

Since the outbreak of the COVID-19 pandemic, videoconferencing has become the default mode of communication in our daily lives at homes, workplaces and schools, and it is likely to remain an important part of our lives in the post-pandemic world. Despite its significance, there has not been any systematic study characterizing the user-perceived performance of existing videoconferencing systems other than anecdotal reports. In this paper, we present a detailed measurement study that compares three major videoconferencing systems: Zoom, Webex and Google Meet. Our study is based on 62 hours’ worth of more than 1.1K videoconferencing sessions, which were created with a mix of emulated videoconferencing clients deployed in the cloud, as well as real mobile devices running from a residential network over two separate periods with nine months apart. We find that the existing videoconferencing systems vary in terms of geographic scope and resource provisioning strategies, which in turns determine streaming lag experienced by users. We also observe that streaming rate can change under different conditions (e.g., available bandwidth, number of users in a session, mobile device status), which affects user-perceived streaming quality. Beyond these findings, our measurement methodology enables reproducible benchmark analysis for any types of comparative or longitudinal study on available videoconferencing systems. [ABSTRACT FROM AUTHOR]

Numerous tracking approaches attempt to improve target representation through target-aware or distractor-aware. However, the unbalanced considerations of target or distractor information make it diffcult for these methods to benefit from the two aspects at the same time. In this paper, we propose a target-distractor aware model with discriminative enhancement learning loss to learn target representation, which can better distinguish the target in complex scenes. Firstly, to enlarge the gap between the target and distractor, we design a discriminative enhancement learning loss. By highlighting the hard negatives that are similar to the target and shrinking the easy negatives that are pure background, the features sensitive to the target or distractor representation can be more conveniently mined. On this basis, we further propose a target-distractor aware model. Unlike existing methods of preference target or distractor, we construct the target-specific feature space by activating the target-sensitive and the distractor-silence feature. Therefore, the appearance model can not only represent the target well but also suppress the background distractor. Finally, the target-distractor aware target representation model is integrated with a Siamese matching network for visual tracking for achieving robust and realtime visual tracking. Extensive experiments are performed on eight tracking benchmarks show that the proposed algorithm achieves favorable performance. [ABSTRACT FROM AUTHOR]

Due to the attenuation and scattering of light by water, there are many quality defects in raw underwater images such as color casts, decreased visibility, reduced contrast, et al.. Many different underwater image enhancement (UIE) algorithms have been proposed to enhance underwater image quality. However, how to fairly compare the performance among UIE algorithms remains a challenging problem. So far, the lack of comprehensive human subjective user study with large-scale benchmark dataset and reliable objective image quality assessment (IQA) metric makes it difficult to fully understand the true performance of UIE algorithms. We in this paper make efforts in both subjective and objective aspects to fill these gaps. Firstly, we construct a new Subjectively-Annotated UIE benchmark Dataset (SAUD) which simultaneously provides real-world raw underwater images, readily available enhanced results by representative UIE algorithms, and subjective ranking scores of each enhanced result. Secondly, we propose an effective No-reference (NR) Underwater Image Quality metric (NUIQ) to automatically evaluate the visual quality of enhanced underwater images. Experiments on the constructed SAUD dataset demonstrate the superiority of our proposed NUIQ metric, achieving higher consistency with subjective rankings than 22 mainstream NR-IQA metrics. The dataset and source code will be made available at https://github.com/yia-yuese/SAUD-Dataset. [ABSTRACT FROM AUTHOR]

This paper generalizes the Attention in Attention (AiA) mechanism, in P. Fang et al., 2019 by employing explicit mapping in reproducing kernel Hilbert spaces to generate attention values of the input feature map. The AiA mechanism models the capacity of building inter-dependencies among the local and global features by the interaction of inner and outer attention modules. Besides a vanilla AiA module, termed linear attention with AiA, two non-linear counterparts, namely, second-order polynomial attention and Gaussian attention, are also proposed to utilize the non-linear properties of the input features explicitly, via the second-order polynomial kernel and Gaussian kernel approximation. The deep convolutional neural network, equipped with the proposed AiA blocks, is referred to as Attention in Attention Network (AiA-Net). The AiA-Net learns to extract a discriminative pedestrian representation, which combines complementary person appearance and corresponding part features. Extensive ablation studies verify the effectiveness of the AiA mechanism and the use of non-linear features hidden in the feature map for attention design. Furthermore, our approach outperforms current state-of-the-art by a considerable margin across a number of benchmarks. In addition, state-of-the-art performance is also achieved in the video person retrieval task with the assistance of the proposed AiA blocks. [ABSTRACT FROM AUTHOR]

Stereo matching is one of the longest-standing problems in computer vision with close to 40 years of studies and research. Throughout the years the paradigm has shifted from local, pixel-level decision to various forms of discrete and continuous optimization to data-driven, learning-based methods. Recently, the rise of machine learning and the rapid proliferation of deep learning enhanced stereo matching with new exciting trends and applications unthinkable until a few years ago. Interestingly, the relationship between these two worlds is two-way. While machine, and especially deep, learning advanced the state-of-the-art in stereo matching, stereo itself enabled new ground-breaking methodologies such as self-supervised monocular depth estimation based on deep networks. In this paper, we review recent research in the field of learning-based depth estimation from single and binocular images highlighting the synergies, the successes achieved so far and the open challenges the community is going to face in the immediate future. [ABSTRACT FROM AUTHOR]

Cognizant of the ease with which softwarized functions can be dynamically scaled according to real time resource requirements, and the fact that multiple services can have common VNFs in their chaining, this paper tackles the problem of cost effective deployment of online services from the perspective of sharing their VNF instances. First, we formally formulate the deployment problem under VNFs sharing. Secondly, given the NP-hard nature of the above problem, we propose a reinforcement learning (RL) algorithm capable of making intelligent placement decisions while considering multiple conflicting costs. Costs of transmission, VNF instantiation or energy consumption, among others. Thanks to the intelligence of the RL algorithm, simulation results show that the performance of the proposed algorithm is within a 14% margin and similar to an optimal solution in terms of request provisioning cost and acceptance ratio, respectively. Moreover, the algorithm results in more than a 20% and a 70% improvement in terms of request deployment cost and time compared to a state-of-the-art algorithm, and up to more than a 40% improvement in terms of cost compared to an algorithm that greedily minimizes the transmission or VNF activation costs. [ABSTRACT FROM AUTHOR]

Accurate numerical modeling of structural components is necessary for both performance-based designs of new structures and performance assessments of existing structures. Steel special concentrically braced frames (SCBFs) represent a large portfolio of steel buildings in areas of high seismicity, especially in North America. Therefore, it is necessary to critically study the accuracy of available numerical models for SCBFs and propose improvements to them, if necessary. This requires forming a database of high-fidelity benchmark experimental results for SCBFs against which models can be compared. This paper presents multielement hybrid simulations on a steel SCBF, as a representation of systems with highly unsymmetrical hysteretic response. A representative buckling brace specimen is developed and verified through reversed-cyclic tests to physically capture the response of SCBFs in the experiments. 2-element and 4-element pseudo-dynamic hybrid simulations are carried out on the reference SCBF using the University of Toronto Ten-Element Hybrid Simulation Platform (UT10). The experimental results from the multielement tests are compared with numerical simulations in detail using the available numerical models, which are observed to predict the response well. [ABSTRACT FROM AUTHOR]

This paper studies joint beamforming problems for an intelligent reflecting surface (IRS)-aided multi-cell multiple-input single-output (MISO) system, and the goal is to maximize the sum rate by jointly optimizing the transmit beamforming vectors at BSs and the reflective beamforming vector at the IRS, subject to the individual maximum transmit power constraints at BSs, and the reflection constraints at the IRS. Due to the formulated optimization problem is highly non-convex, we propose an alternating optimization (AO) algorithm based on successive convex approximation (SCA) such that the transmit and reflective beamforming vectors can be optimized alternately. We further consider the SINR balancing beamforming design scheme by maximizing the minimum SINR among all users to enhance the fairness among users, in which the transmit and reflective beamforming vectors are optimized in an alternating manner. The transmit beamforming vectors are optimized by the second-order-cone programming (SOCP) based on bisection method and the reflective beamforming vector is updated based on the technique of semidefinite relaxation (SDR). Simulation results show that the two proposed algorithms considerably outperform the benchmark zero-forcing (ZF) scheme. Moreover, the AO algorithm based on SCA has good communication performance than the other two schemes. And the AO algorithm based on bisection search guarantees the fairness for all users. [ABSTRACT FROM AUTHOR]

Many wireless networks, including 5G NR (New Radio) and future beyond 5G cellular systems, are expected to operate on multiple frequency bands. This paper considers the band assignment (BA) problem in dual-band systems, where the basestation (BS) chooses one of the two available frequency bands (centimeter-wave and millimeter-wave bands) to communicate with the user equipment (UE). While the millimeter-wave band might offer higher data rate, there is a significant probability of outage during which the communication should be carried on the (more reliable) centimeter-wave band. With mobility, the BA can be perceived as a sequential problem, where the BS uses previously observed information to predict the best band for a future time step. We formulate the BA as a binary classification problem and propose supervised Machine Learning (ML) solutions. We study the problem when both the BS and the UE use (i) omnidirectional antennas and (ii) both use directional antennas. In the omnidirectional case, we derive analytical benchmark solutions based on the Gaussian Process (GP) assumption for the inter-band shadow fading. In the directional case, where the labeling is shown to be complex, we propose an efficient labeling approach based on the Viterbi Algorithm (VA). We compare the performances for two channel models: (i) a stochastic channel and (ii) a ray-tracing based channel. [ABSTRACT FROM AUTHOR]

Ophthalmologists have used fundus images to screen and diagnose eye diseases. However, different equipments and ophthalmologists pose large variations to the quality of fundus images. Low-quality (LQ) degraded fundus images easily lead to uncertainty in clinical screening and generally increase the risk of misdiagnosis. Thus, real fundus image restoration is worth studying. Unfortunately, real clinical benchmark has not been explored for this task so far. In this paper, we investigate the real clinical fundus image restoration problem. Firstly, We establish a clinical dataset, Real Fundus (RF), including 120 low- and high-quality (HQ) image pairs. Then we propose a novel Transformer-based Generative Adversarial Network (RFormer) to restore the real degradation of clinical fundus images. The key component in our network is the Window-based Self-Attention Block (WSAB) which captures non-local self-similarity and long-range dependencies. To produce more visually pleasant results, a Transformer-based discriminator is introduced. Extensive experiments on our clinical benchmark show that the proposed RFormer significantly outperforms the state-of-the-art (SOTA) methods. In addition, experiments of downstream tasks such as vessel segmentation and optic disc/cup detection demonstrate that our proposed RFormer benefits clinical fundus image analysis and applications. [ABSTRACT FROM AUTHOR]

Image compression is one of the most fundamental techniques and commonly used applications in the image and video processing field. Earlier methods built a well-designed pipeline, and efforts were made to improve all modules of the pipeline by handcrafted tuning. Later, tremendous contributions were made, especially when data-driven methods revitalized the domain with their excellent modeling capacities and flexibility in incorporating newly designed modules and constraints. Despite great progress, a systematic benchmark and comprehensive analysis of end-to-end learned image compression methods are lacking. In this paper, we first conduct a comprehensive literature survey of learned image compression methods. The literature is organized based on several aspects to jointly optimize the rate-distortion performance with a neural network, i.e., network architecture, entropy model and rate control. We describe milestones in cutting-edge learned image-compression methods, review a broad range of existing works, and provide insights into their historical development routes. With this survey, the main challenges of image compression methods are revealed, along with opportunities to address the related issues with recent advanced learning methods. This analysis provides an opportunity to take a further step towards higher-efficiency image compression. By introducing a coarse-to-fine hyperprior model for entropy estimation and signal reconstruction, we achieve improved rate-distortion performance, especially on high-resolution images. Extensive benchmark experiments demonstrate the superiority of our model in rate-distortion performance and time complexity on multi-core CPUs and GPUs. [ABSTRACT FROM AUTHOR]

Small-sized unmanned surface vehicles (USV) are coastal water devices with a broad range of applications such as environmental control and surveillance. A crucial capability for autonomous operation is obstacle detection for timely reaction and collision avoidance, which has been recently explored in the context of camera-based visual scene interpretation. Owing to curated datasets, substantial advances in scene interpretation have been made in a related field of unmanned ground vehicles. However, the current maritime datasets do not adequately capture the complexity of real-world USV scenes and the evaluation protocols are not standardised, which makes cross-paper comparison of different methods difficult and hinders the progress. To address these issues, we introduce a new obstacle detection benchmark MODS, which considers two major perception tasks: maritime object detection and the more general maritime obstacle segmentation. We present a new diverse maritime evaluation dataset containing approximately 81k stereo images synchronized with an on-board IMU, with over 60k objects annotated. We propose a new obstacle segmentation performance evaluation protocol that reflects the detection accuracy in a way meaningful for practical USV navigation. Nineteen recent state-of-the-art object detection and obstacle segmentation methods are evaluated using the proposed protocol, creating a benchmark to facilitate development of the field. The proposed dataset, as well as evaluation routines, are made publicly available at vicos.si/resources. [ABSTRACT FROM AUTHOR]

Privacy-preserving deep learning with homomorphic encryption (HE) is a novel and promising research area aimed at designing deep learning solutions that operate while guaranteeing the privacy of user data. Designing privacy-preserving deep learning solutions requires one to completely rethink and redesign deep learning models and algorithms to match the severe technological and algorithmic constraints of HE. This paper provides an introduction to this complex research area as well as a methodology for designing privacy-preserving convolutional neural networks (CNNs). This methodology was applied to the design of a privacy-preserving version of the well-known LeNet-1 CNN, which was successfully operated on two benchmark datasets for image classification. Furthermore, this paper details and comments on the research challenges and software resources available for privacy-preserving deep learning with HE. [ABSTRACT FROM AUTHOR]

In order to reduce the gap between the laboratory environment and actual use in daily life of human-machine interaction based on surface electromyogram (sEMG) intent recognition, this paper presents a benchmark dataset of sEMG in non-ideal conditions (SeNic). The dataset mainly consists of 8-channel sEMG signals, and electrode shifts from an 3D-printed annular ruler. A total of 36 subjects participate in our data acquisition experiments of 7 gestures in non-ideal conditions, where non-ideal factors of 1) electrode shifts, 2) individual difference, 3) muscle fatigue, 4) inter-day difference, and 5) arm postures are elaborately involved. Signals of sEMG are validated first in temporal and frequency domains. Results of recognizing gestures in ideal conditions indicate the high quality of the dataset. Adverse impacts in non-ideal conditions are further revealed in the amplitudes of these data and recognition accuracies. To be concluded, SeNic is a benchmark dataset that introduces several non-ideal factors which often degrade the robustness of sEMG-based systems. It could be used as a freely available dataset and a common platform for researchers in the sEMG-based recognition community. The benchmark dataset SeNic are available online via the website (https://github.com/bozhubo/SeNic and https://gitee.com/bozhubo/SeNic). [ABSTRACT FROM AUTHOR]

Weakly-supervised video object segmentation is an emerging video task to track and segment the target given a simple bounding box label, which requires the method to fully catch and utilize the target information. Most existing approaches only rely on the guidance of a single frame and ignore the interaction between different frames when gathering information, making them hard to achieve reliable target representation. In this paper, we propose to capture the temporal dependencies and gather information from multiple frames through bilateral temporal re-aggregation. We explore three schemes to build the aggregation: 1) a two-stage re-aggregation mechanism is applied to provide target prior to the current frame, which obtains more valid feature matching and information aggregation; 2) a query-memory bilateral aggregation module is proposed to aggregate features from an unlimited amount of past frames and enable the mutual perception between different frames to validate the gathered information; 3) we guide the learning of aggregation modules through a novel cross-task representation distillation, transferring the knowledge from a semi-supervised model to our weakly-supervised model without increasing the inference latency. These schemes collaboratively build an efficient and competent aggregation process, thus we can fully exploit the video context to make the inference. Experimental results on four benchmarks show that our method achieves superior performance than previous methods and still maintains the efficiency ($e.g$., overall scores of 70.4% and 72.5% on the YouTube-VOS and DAVIS 2017 validation sets, respectively). [ABSTRACT FROM AUTHOR]