Search

Transfer learning enables to re-use knowledge learned on a source task to help learning a target task. A simple form of transfer learning is common in current state-of-the-art computer vision models, i.e., pre-training a model for image classification on the ILSVRC dataset, and then fine-tune on any target task. However, previous systematic studies of transfer learning have been limited and the circumstances in which it is expected to work are not fully understood. In this paper we carry out an extensive experimental exploration of transfer learning across vastly different image domains (consumer photos, autonomous driving, aerial imagery, underwater, indoor scenes, synthetic, close-ups) and task types (semantic segmentation, object detection, depth estimation, keypoint detection). Importantly, these are all complex, structured output tasks types relevant to modern computer vision applications. In total we carry out over 2000 transfer learning experiments, including many where the source and target come from different image domains, task types, or both. We systematically analyze these experiments to understand the impact of image domain, task type, and dataset size on transfer learning performance. Our study leads to several insights and concrete recommendations: (1) for most tasks there exists a source which significantly outperforms ILSVRC’12 pre-training; (2) the image domain is the most important factor for achieving positive transfer; (3) the source dataset should include the image domain of the target dataset to achieve best results; (4) at the same time, we observe only small negative effects when the image domain of the source task is much broader than that of the target; (5) transfer across task types can be beneficial, but its success is heavily dependent on both the source and target task types. [ABSTRACT FROM AUTHOR]

Transfer learning enables to re-use knowledge learned on a source task to help learning a target task. A simple form of transfer learning is common in current state-of-the-art computer vision models, i.e., pre-training a model for image classification on the ILSVRC dataset, and then fine-tune on any target task. However, previous systematic studies of transfer learning have been limited and the circumstances in which it is expected to work are not fully understood. In this paper we carry out an extensive experimental exploration of transfer learning across vastly different image domains (consumer photos, autonomous driving, aerial imagery, underwater, indoor scenes, synthetic, close-ups) and task types (semantic segmentation, object detection, depth estimation, keypoint detection). Importantly, these are all complex, structured output tasks types relevant to modern computer vision applications. In total we carry out over 2000 transfer learning experiments, including many where the source and target come from different image domains, task types, or both. We systematically analyze these experiments to understand the impact of image domain, task type, and dataset size on transfer learning performance. Our study leads to several insights and concrete recommendations: (1) for most tasks there exists a source which significantly outperforms ILSVRC’12 pre-training; (2) the image domain is the most important factor for achieving positive transfer; (3) the source dataset should include the image domain of the target dataset to achieve best results; (4) at the same time, we observe only small negative effects when the image domain of the source task is much broader than that of the target; (5) transfer across task types can be beneficial, but its success is heavily dependent on both the source and target task types. [ABSTRACT FROM AUTHOR]

To improve the viewer’s Quality of Experience (QoE) and optimize computer graphics applications, 3D model quality assessment (3D-QA) has become an important task in the multimedia area. Point cloud and mesh are the two most widely used digital representation formats of 3D models, the visual quality of which is quite sensitive to lossy operations like simplification and compression. Therefore, many related studies such as point cloud quality assessment (PCQA) and mesh quality assessment (MQA) have been carried out to measure the visual quality of distorted 3D models. However, most previous studies utilize full-reference (FR) metrics, which indicates they can not predict the quality level in the absence of the reference 3D model. Furthermore, few 3D-QA metrics consider color information, which significantly restricts their effectiveness and scope of application. In this paper, we propose a no-reference (NR) quality assessment metric for colored 3D models represented by both point cloud and mesh. First, we project the 3D models from 3D space into quality-related geometry and color feature domains. Then, the 3D natural scene statistics (3D-NSS) and entropy are utilized to extract quality-aware features. Finally, a support vector regression (SVR) model is employed to regress the quality-aware features into visual quality scores. Our method is validated on the colored point cloud quality assessment database (SJTU-PCQA), the Waterloo point cloud assessment database (WPC), and the colored mesh quality assessment database (CMDM). The experimental results show that the proposed method outperforms most compared NR 3D-QA metrics with competitive computational resources and greatly reduces the performance gap with the state-of-the-art FR 3D-QA metrics. The code of the proposed model is publicly available now at https://github.com/zzc-1998/NR-3DQA. [ABSTRACT FROM AUTHOR]

Image smoothing is a fundamental procedure in applications of both computer vision and graphics. The required smoothing properties can be different or even contradictive among different tasks. Nevertheless, the inherent smoothing nature of one smoothing operator is usually fixed and thus cannot meet the various requirements of different applications. In this paper, we first introduce the truncated Huber penalty function which shows strong flexibility under different parameter settings. A generalized framework is then proposed with the introduced truncated Huber penalty function. When combined with its strong flexibility, our framework is able to achieve diverse smoothing natures where contradictive smoothing behaviors can even be achieved. It can also yield the smoothing behavior that can seldom be achieved by previous methods, and superior performance is thus achieved in challenging cases. These together enable our framework capable of a range of applications and able to outperform the state-of-the-art approaches in several tasks, such as image detail enhancement, clip-art compression artifacts removal, guided depth map restoration, image texture removal, etc. In addition, an efficient numerical solution is provided and its convergence is theoretically guaranteed even the optimization framework is non-convex and non-smooth. A simple yet effective approach is further proposed to reduce the computational cost of our method while maintaining its performance. The effectiveness and superior performance of our approach are validated through comprehensive experiments in a range of applications. Our code is available at https://github.com/wliusjtu/Generalized-Smoothing-Framework . [ABSTRACT FROM AUTHOR]

Many computer vision applications rely on feature detection and description, hence the need for computationally efficient and robust 4D light field (LF) feature detectors and descriptors. In this paper, we propose a novel light field feature descriptor based on the Fourier disparity layer representation, for light field imaging applications. After the Harris feature detection in a scale-disparity space, the proposed feature descriptor is then extracted using a circular neighborhood rather than a square neighborhood. It is shown to yield more accurate feature matching, compared with the LiFF LF feature, with a lower computational complexity. In order to evaluate the feature matching performance with the proposed descriptor, we generated a synthetic stereo LF dataset with ground truth matching points. Experimental results with synthetic and real-world dataset show that our solution outperforms existing methods in terms of both feature detection robustness and feature matching accuracy. [ABSTRACT FROM AUTHOR]

Homography estimation is an important task in computer vision applications, such as image stitching, video stabilization, and camera calibration. Traditional homography estimation methods heavily depend on the quantity and distribution of feature correspondences, leading to poor robustness in low-texture scenes. The learning solutions, on the contrary, try to learn robust deep features but demonstrate unsatisfying performance in the scenes with low overlap rates. In this paper, we address these two problems simultaneously by designing a contextual correlation layer (CCL). The CCL can efficiently capture the long-range correlation within feature maps and can be flexibly used in a learning framework. In addition, considering that a single homography can not represent the complex spatial transformation in depth-varying images with parallax, we propose to predict multi-grid homography from global to local. Moreover, we equip our network with a depth perception capability, by introducing a novel depth-aware shape-preserved loss. Extensive experiments demonstrate the superiority of our method over state-of-the-art solutions in the synthetic benchmark dataset and real-world dataset. The codes and models will be available at https://github.com/nie-lang/Multi-Grid-Deep-Homography. [ABSTRACT FROM AUTHOR]

Approximate Nearest Neighbor(ANN) search is the core problem in many large-scale machine learning and computer vision applications such as multimodal retrieval. Hashing is becoming increasingly popular, since it can provide efficient similarity search and compact data representations suitable for handling such large-scale ANN search problems. Most hashing algorithms concentrate on learning more effective projection functions. However, the accuracy loss in the quantization step has been ignored and barely studied. In this paper, we analyse the importance of various projected dimensions, distribute them into several groups and quantize them with two types of values which can both better preserve the neighborhood structure among data. One is Variable Integer-based Quantization (VIQ) that quantizes each projected dimension with integer values. The other is Variable Codebook-based Quantization (VCQ) that quantizes each projected dimension with corresponding codebook values. We conduct experiments on five common public data sets containing up to one million vectors. The results show that the proposed VCQ and VIQ algorithms can both achieve much higher accuracy than state-of-the-art quantization methods. Furthermore, although VCQ performs better than VIQ, ANN search with VIQ provides much higher search efficiency. [ABSTRACT FROM AUTHOR]

Low-light images suffer from low contrast and unclear details, which not only reduces the available information for humans but limits the application of computer vision algorithms. Among the existing enhancement techniques, Retinex-based and learning-based methods are under the spotlight today. In this paper, we bridge the gap between the two methods. First, we propose a novel “generative” strategy for Retinex decomposition, by which the decomposition is cast as a generative problem. Second, based on the strategy, a unified deep framework is proposed to estimate the latent components and perform low-light image enhancement. Third, our method can weaken the coupling relationship between the two components while performing Retinex decomposition. Finally, the RetinexDIP performs Retinex decomposition without any external images, and the estimated illumination can be easily adjusted and is used to perform enhancement. The proposed method is compared with ten state-of-the-art algorithms on seven public datasets, and the experimental results demonstrate the superiority of our method. Code is available at: https://github.com/zhaozunjin/RetinexDIP. [ABSTRACT FROM AUTHOR]

This paper introduces a generalized approach of the linear combination of configuration field (LCCF) method. The technique developed in this paper provides a way to identify the temporal profile of electromagnetic sources satisfying particular properties for a given target (fields, voltage, or current). Here, we focus on identifying voltage signals that would lead to a specified voltage over a time interval at some points of a network. To experimentally apply the same idea, we propose two possibilities to decrease the maximum frequencies of the voltage signals. The LCCF method is then applied to bring a software correction to a lossy transmission-line network presenting soft or hard defects. [ABSTRACT FROM AUTHOR]

Temperature rise simulation deals with different physical domains describing heat generation and heat dissipation effects. Cigré working group A3.36 has conducted a study on temperature rise simulation of switchgear for distribution and transmission level. The working group built up new, manufacturer independent devices. Members of the working group simulated the temperature rise in advance to the tests, which were performed at laboratories of universities, as independent parties. In order to assess the simulation accuracy, the simulation results are compared to each other and to the measured results. This paper summarizes the main findings from the medium voltage device. The study has shown that when simulating the temperature rise on a completely new design, with reliable input values of contact resistances and emissivity coefficients, an accuracy of about 25% could be expected. Primary reason seems to be based on the complexity of simulating the heat dissipation mechanisms. To get a more accurate result, verification and validation is needed in order to adjust the parameters determining the heat dissipation. This study has also illustrated that it requires good knowledge (especially of the radiation) in order to adjust these parameters properly. [ABSTRACT FROM AUTHOR]

We have witnessed a growing interest in video salient object detection (VSOD) techniques in today’s computer vision applications. In contrast with temporal information (which is still considered a rather unstable source thus far), the spatial information is more stable and ubiquitous, thus it could influence our vision system more. As a result, the current main-stream VSOD approaches have inferred and obtained their saliency primarily from the spatial perspective, still treating temporal information as subordinate. Although the aforementioned methodology of focusing on the spatial aspect is effective in achieving a numeric performance gain, it still has two critical limitations. First, to ensure the dominance by the spatial information, its temporal counterpart remains inadequately used, though in some complex video scenes, the temporal information may represent the only reliable data source, which is critical to derive the correct VSOD. Second, both spatial and temporal saliency cues are often computed independently in advance and then integrated later on, while the interactions between them are omitted completely, resulting in saliency cues with limited quality. To combat these challenges, this paper advocates a novel spatiotemporal network, where the key innovation is the design of its temporal unit. Compared with other existing competitors (e.g., convLSTM), the proposed temporal unit exhibits an extremely lightweight design that does not degrade its strong ability to sense temporal information. Furthermore, it fully enables the computation of temporal saliency cues that interact with their spatial counterparts, ultimately boosting the overall VSOD performance and realizing its full potential towards mutual performance improvement for each. The proposed method is easy to implement yet still effective, achieving high-quality VSOD at 50 FPS in real-time applications. [ABSTRACT FROM AUTHOR]

Shadow removal can significantly improve the image visual quality and has many applications in computer vision. Deep learning methods based on CNNs have become the most effective approach for shadow removal by training on either paired data, where both the shadow and underlying shadow-free versions of an image are known, or unpaired data, where shadow and shadow-free training images are totally different with no correspondence. In practice, CNN training on unpaired data is more preferred given the easiness of training data collection. In this paper, we present a new Lightness-Guided Shadow Removal Network (LG-ShadowNet) for shadow removal by training on unpaired data. In this method, we first train a CNN module to compensate for the lightness and then train a second CNN module with the guidance of lightness information from the first CNN module for final shadow removal. We also introduce a loss function to further utilise the colour prior of existing data. Extensive experiments on widely used ISTD, adjusted ISTD and USR datasets demonstrate that the proposed method outperforms the state-of-the-art methods with training on unpaired data. [ABSTRACT FROM AUTHOR]

Superpixel segmentation, as a central image processing task, has many applications in computer vision and computer graphics. Boundary alignment and shape compactness are leading indicators to evaluate a superpixel segmentation algorithm. Furthermore, convexity can make superpixels reflect more geometric structures in images and provide a more concise over-segmentation result. In this paper, we consider generating convex and compact superpixels while satisfying the constraints of adhering to the boundary as far as possible. We formulate the new superpixel segmentation into an edge-constrained centroidal power diagram (ECCPD) optimization problem. In the implementation, we optimize the superpixel configurations by repeatedly performing two alternative operations, which include site location updating and weight updating through a weight function defined by image features. Compared with existing superpixel methods, our method can partition an image into fully convex and compact superpixels with better boundary adherence. Extensive experimental results show that our approach outperforms existing superpixel segmentation methods in boundary alignment and compactness for generating convex superpixels. [ABSTRACT FROM AUTHOR]

In this paper, we study the problem of exploiting parallelism in a hard real-time streaming application modeled as an acyclic synchronous data flow (SDF) graph and scheduled on a heterogeneous multiprocessor system-on-chip platform to alleviate the capacity fragmentation due to partitioned scheduling algorithms and reduce the number of required processors when a throughput requirement is satisfied. As the main contribution in this paper, we propose a method to determine a replication factor for each task in an acyclic SDF graph such that by distributing the workloads among more parallel tasks with lower utilization in the obtained transformed graph, the left capacity on the processors can be efficiently exploited, hence reducing the number of required processors. The experimental results, on a set of real-life streaming applications, demonstrate that our approach can reduce the minimum number of processors required to schedule an application and considerably improve the memory requirements and application latency compared to related approaches while meeting the same throughput constraint. [ABSTRACT FROM AUTHOR]

This paper introduces a family of quasi-linear discriminants that outperform current large-margin methods in sliding window visual object detection and open set recognition tasks. In these applications, the classification problems are both numerically imbalanced – positive (object class) training and test windows are much rarer than negative (non-class) ones – and geometrically asymmetric – the positive samples typically form compact, visually-coherent groups while negatives are much more diverse, including anything at all that is not a well-centered sample from the target class. For such tasks, there is a need for discriminants whose decision regions focus on tightly circumscribing the positive class, while still taking account of negatives in zones where the two classes overlap. To this end, we propose a family of quasi-linear “polyhedral conic” discriminants whose positive regions are distorted $L_1$ L 1 or $L_2$ L 2 balls. In addition, we also integrated the proposed classification loss into deep neural networks so that both the features and classifier can be learned simultaneously end-to-end fashion to improve the classification accuracies. The methods have properties and run-time complexities comparable to linear Support Vector Machines (SVMs), and they can be trained from either binary or positive-only samples using constrained quadratic programs related to SVMs. Our experiments show that they significantly outperform linear SVMs, deep neural networks using softmax loss function and existing one-class discriminants on a wide range of object detection, face verification, open set recognition and conventional closed-set classification tasks. [ABSTRACT FROM AUTHOR]

Removing the undesired reflections from images taken through the glass is of broad application to various computer vision tasks. Non-learning based methods utilize different handcrafted priors such as the separable sparse gradients caused by different levels of blurs, which often fail due to their limited description capability to the properties of real-world reflections. In this paper, we propose a network with the feature-sharing strategy to tackle this problem in a cooperative and unified framework, by integrating image context information and the multi-scale gradient information. To remove the strong reflections existed in some local regions, we propose a statistic loss by considering the gradient level statistics between the background and reflections. Our network is trained on a new dataset with 3250 reflection images taken under diverse real-world scenes. Experiments on a public benchmark dataset show that the proposed method performs favorably against state-of-the-art methods. [ABSTRACT FROM AUTHOR]

The emerging large-scale multi-socket systems make the need for more sophisticated large-scale coherence management of necessity. Directory-based coherence has been an ad hoc solution and a clear candidate for large-scale shared-memory systems. A vanilla directory design, however, suffers from inefficient use of storage to keep coherence metadata, resulting in a high storage overhead for large-scale systems. In this paper, we propose a dynamic multi-grain directory for large multi-socket systems. The idea is to track coherence of regions of different sizes which requires storing much less information in the directory than having a directory entry per each data block. It dynamically refines granularity according to the application phase and therefore tracks coherence information for regions of varying sizes. The results show that the proposal allows to reduce the directory storage by an order of magnitude, while the loss of precision does not cause performance penalty. The paper demonstrates that different applications and different application phases have different requirements for the region size. Performance results are compared against two state-of-the-art multi-grain directories and it is the one that obtains the better results. [ABSTRACT FROM AUTHOR]

Securing cyber-physical systems (CPS) and Internet of Things (IoT) systems requires the identification of how interdependence among existing atomic vulnerabilities may be exploited by an adversary to stitch together an attack that can compromise the system. Therefore, accurate attack graphs play a significant role in systems security. A manual construction of the attack graphs is tedious and error-prone, this paper proposes a model-checking-based automated attack graph generator and visualizer (A2G2V). The proposed A2G2V algorithm uses existing model-checking tools, an architecture description tool, and our own code to generate an attack graph that enumerates the set of all possible sequences in which atomic-level vulnerabilities can be exploited to compromise system security. The architecture description tool captures a formal representation of the networked system, its atomic vulnerabilities, their pre-and post-conditions, and security property of interest. A model-checker is employed to automatically identify an attack sequence in the form of a counterexample. Our own code integrated with the model-checker parses the counterexamples, encodes those for specification relaxation, and iterates until all attack sequences are revealed. Finally, a visualization tool has also been incorporated with A2G2V to generate a graphical representation of the generated attack graph. The results are illustrated through application to computer as well as control (SCADA) networks. [ABSTRACT FROM AUTHOR]

Superpixels are widely used in computer vision applications, as they conserve the running costs of subsequent processing while preserving the original performance. In most of the existing algorithms, the boundary adherence and the compactness of superpixels are necessarily inter-inhibitive because the color/gradient information is balanced against the position constraints, and the set criteria define all pixels indiscriminately. In this paper, we present a two-phase superpixel segmentation method based on the watershed transformation. After designing a new approach for calculating the flooding priority, we propose a new strategy with two distinct criteria for global and local refinement of the boundary pixels. These criteria reduce the compromise between the boundary adherence and compactness. Unlike the indiscriminate standards, our method applies different treatments to pixels in different environments, preserving the color homogeneity in content-rich areas while improving the regularity of the superpixels in content-plain regions. The superior accuracy and computing time of our proposed method are verified in comparison experiments with several state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Online image hashing aims to update hash functions on-the-fly along with newly arriving data streams, which has found broad applications in computer vision and beyond. To this end, most existing methods update hash functions simply using discrete labels or pairwise similarity to explore intra-class relationships, which, however, often deteriorates search performance when facing a domain gap or semantic shift. One reason is that they ignore the particular semantic relationships among different classes, which should be taken into account in updating hash functions. Besides, the common characteristics between the label vectors (can be regarded as a sort of binary codes) and to-be-learned binary hash codes have left unexploited. In this paper, we present a novel online hashing method, termed Similarity Preserving Linkage Hashing (SPLH), which not only utilizes pairwise similarity to learn the intra-class relationships, but also fully exploits a latent linkage space to capture the inter-class relationships and the common characteristics between label vectors and to-be-learned hash codes. Specifically, SPLH first maps the independent discrete label vectors and binary hash codes into a linkage space, through which the relative semantic distance between data points can be assessed precisely. As a result, the pairwise similarities within the newly arriving data stream are exploited to learn the latent semantic space to benefit binary code learning. To learn the model parameters effectively, we further propose an alternating optimization algorithm. Extensive experiments conducted on three widely-used datasets demonstrate the superior performance of SPLH over several state-of-the-art online hashing methods. [ABSTRACT FROM AUTHOR]

The guided filter and its subsequent derivatives have been widely employed in many image processing and computer vision applications primarily brought about by their low complexity and good edge-preservation properties. Despite this success, the different variants of the guided filter are unable to handle more aggressive filtering strengths leading to the manifestation of “detail halos”. At the same time, these existing filters perform poorly when the input and guide images have structural inconsistencies. In this paper, we demonstrate that these limitations are due to the guided filter operating as a variable-strength locally-isotropic filter that, in effect, acts as a weak anisotropic filter on the image. Our analysis shows that this behaviour stems from the use of unweighted averaging in the final steps of guided filter variants including the adaptive guided filter (AGF), weighted guided image filter (WGIF), and gradient-domain guided image filter (GGIF). We propose a novel filter, the Anisotropic Guided Filter (AnisGF), that utilises weighted averaging to achieve maximum diffusion while preserving strong edges in the image. The proposed weights are optimised based on the local neighbourhood variances to achieve strong anisotropic filtering while preserving the low computational cost of the original guided filter. Synthetic tests show that the proposed method addresses the presence of detail halos and the handling of inconsistent structures found in previous variants of the guided filter. Furthermore, experiments in scale-aware filtering, detail enhancement, texture removal, and chroma upsampling demonstrate the improvements brought about by the technique. [ABSTRACT FROM AUTHOR]

3-D convolutional neural networks (3-D CNNs) are used efficiently in many computer vision applications. Most previous work in this area has concentrated only on design and optimization of accelerators for 2-D CNNs, with few attempts having been made to accelerate 3-D CNNs on FPGA. We find the acceleration of 3-D CNNs on FPGA to be challenging due to their high computational complexity and storage demands. More importantly, although the computational patterns of 2-D and 3-D CNNs are analogous, the conventional approaches that have been adopted for acceleration of 2-D CNNs may be unfit for 3-D CNN acceleration. In this paper, in order to accelerate 2-D and 3-D CNNs using a uniform framework, we first propose a uniform template-based architecture that uses templates based on the Winograd algorithm to ensure the rapid development of 2-D and 3-D CNN accelerators. Then, with the aim of efficiently mapping all layers of 2-D /3-D CNNs onto a pipelined accelerator, techniques are developed to improve the throughput and computational efficiency of the accelerator, including layer fusion, layer clustering, and workload-balancing scheme. Finally, we demonstrate the effectiveness of the deep pipelined architecture by accelerating real-life 2-D and 3-D CNNs on the state-of-the-art FPGA platform. On VCU118, we achieve 3.7 TOPS for VGG-16, which outperforms state-of-the-art FPGA-based CNN accelerators. Comparisons with CPU and GPU solutions demonstrate that our implementation of 3-D CNN achieves gains of up to $17.8\times $ and $64.2\times $ in performance and energy relative to a CPU solution, and a $5.0\times $ energy efficiency gain over a GPU solution. [ABSTRACT FROM AUTHOR]

Color and color differences are critical aspects in many image processing and computer vision applications. A paradigmatic example is object segmentation, where color distances can greatly influence the performance of the algorithms. Metrics for color difference have been proposed in the literature, including the definition of standards such as CIEDE2000, which quantifies the change in visual perception of two given colors. This standard has been recommended for industrial computer vision applications, but the benefits of its application have been impaired by the complexity of the formula. This paper proposes a new strategy that improves the usability of the CIEDE2000 metric when a maximum acceptable distance can be imposed. We argue that, for applications where a maximum value, above which colors are considered to be different, can be established, then it is possible to reduce the amount of calculations of the metric, by preemptively analyzing the color features. This methodology encompasses the benefits of the metric while overcoming its computational limitations, thus broadening the range of applications of CIEDE2000 in both the computer vision algorithms and computational resource requirements. [ABSTRACT FROM AUTHOR]

Finding the informative subspaces of high-dimensional datasets is at the core of numerous applications in computer vision, where spectral-based subspace clustering is arguably the most widely studied method due to its strong empirical performance. Such algorithms first compute an affinity matrix to construct a self-representation for each sample using other samples as a dictionary. Sparsity and connectivity of the self-representation play important roles in effective subspace clustering. However, simultaneous optimization of both factors is difficult due to their conflicting nature, and most existing methods are designed to address only one factor. In this paper, we propose a post-processing technique to optimize both sparsity and connectivity by finding good neighbors. Good neighbors induce key connections among samples within a subspace and not only have large affinity coefficients but are also strongly connected to each other. We reassign the coefficients of the good neighbors and eliminate other entries to generate a new coefficient matrix. We show that the few good neighbors can effectively recover the subspace, and the proposed post-processing step of finding good neighbors is complementary to most existing subspace clustering algorithms. Experiments on five benchmark datasets show that the proposed algorithm performs favorably against the state-of-the-art methods with negligible additional computation cost. [ABSTRACT FROM AUTHOR]

Robust visual tracking is one of the most challenging problems in computer vision applications. However, the limited training data and the computational complexity have severely affected tracking performance. In this paper, we propose an auto-encoder pair model for visual tracking which is composed of source domain network and target domain network to help a more accurate localization. We adopt the dense circular samples of the object state to increase the number of training samples and prevent model overfitting. Meanwhile, a difference regularization term is also introduced into our framework to penalize the large appearance variations of the object in two domains. The alternating optimization is used to solve the optimization problems. Furthermore, our method alleviates the model update problem and improves the tracking speed by using long-term and short-term updating scheme. In addition, the target domain filter is updated by introducing the updated source domain filter to avoid the object drift. Comprehensive experiments on some challenging benchmarks demonstrate that our approach concurrently improves both tracking accuracy and speed. [ABSTRACT FROM AUTHOR]

Radial distortion, which severely hinders object detection and semantic recognition, frequently exists in images captured using a wide-angle lens. Correction of this distortion of images is crucial in many computer vision applications. In this paper, we present distortion rectification generative adversarial network (DR-GAN), a conditional generative adversarial network (GAN) for automatic radial DR. To the best of our knowledge, this is the first end-to-end trainable adversarial framework for radial distortion rectification. The DR-GAN trained using the proposed low-to-high perceptual loss learns the mapping relation between different structural images rather than estimating multifarious distortion parameters, while also realizing label-free training and one-stage rectification. As a benefit of one-stage rectification, the proposed method is extremely fast with the completion of rectification in real time. This is approximately 22 times faster than the state-of-the-art methods. The experimental results show that the DR-GAN achieves an excellent performance in both quantitative measure (PSNR and SSIM) and visual qualitative appearance. [ABSTRACT FROM AUTHOR]

Correctly matching feature points in a pair of images is an important preprocessing step for many computer vision applications. In this paper we propose an efficient method for estimating the number of correct matches without explicitly computing them. To this end, we propose to analyze the set of matches using the spatial order of the features, as projected to the $x$ x -axis of the image. The set of features in each image is thus represented by a sequence, and analyzed using the Kendall and Spearman Footrule distance metrics between permutations. This result is interesting in its own right. Moreover, we demonstrate three useful applications of our method: (i) a new halting condition for RANSAC based epipolar geometry estimation methods, (ii) discarding spatially unrelated image pairs in the Structure-from-Motion pipeline, and (iii) computing the probability that a given match is correct based on the rank of the features within the sequences. Our experiments on a large number of synthetic and real data demonstrate the effectiveness of our method. For example, the running time of the image matching stage in the Structure-from-Motion pipeline may be reduced by about 90 percent while preserving about 85 percent of the image pairs with spatial overlap. [ABSTRACT FROM AUTHOR]

In recent years, the widespread use of deep neural networks (DNNs) has facilitated great improvements in performance for computer vision tasks like image classification and object recognition. In most realistic computer vision applications, an input image undergoes some form of image distortion such as blur and additive noise during image acquisition or transmission. Deep networks trained on pristine images perform poorly when tested on such distortions. In this paper, we evaluate the effect of image distortions like Gaussian blur and additive noise on the activations of pre-trained convolutional filters. We propose a metric to identify the most noise susceptible convolutional filters and rank them in order of the highest gain in classification accuracy upon correction. In our proposed approach called DeepCorrect, we apply small stacks of convolutional layers with residual connections at the output of these ranked filters and train them to correct the worst distortion affected filter activations, while leaving the rest of the pre-trained filter outputs in the network unchanged. Performance results show that applying DeepCorrect models for common vision tasks like image classification (ImageNet), object recognition (Caltech-101, Caltech-256), and scene classification (SUN-397), significantly improves the robustness of DNNs against distorted images and outperforms other alternative approaches. [ABSTRACT FROM AUTHOR]

Automatic target recognition (ATR) is an important part for many computer vision applications. Despite the extensive research which has been carried out in this area for many years, there is no ATR system which performs well on all applications. Recently, different object recognition frameworks have been proposed which yield a high performance in baseline databases. However, our experiments showed that they can fail in real-world scenarios, when dealing with a limited number of data samples. In this paper, we propose a new ATR system, based on deep convolutional neural network (DCNN), to detect the targets in forward looking infrared (FLIR) scenes and recognize their classes. In our proposed ATR framework, a fully convolutional network is trained to map the input FLIR imagery data to a fixed stride correspondingly-sized target score map. The potential targets are identified by applying a threshold on the target score map. Finally, the corresponding regions centered at these target points are fed to a DCNN to classify them into different target types while at the same time rejecting the false alarms. The proposed architecture achieves a significantly better performance in comparison with that of the state-of-the-art methods on two large FLIR image databases. [ABSTRACT FROM AUTHOR]

Low-rank modeling has many important applications in computer vision and machine learning. While the matrix rank is often approximated by the convex nuclear norm, the use of nonconvex low-rank regularizers has demonstrated better empirical performance. However, the resulting optimization problem is much more challenging. Recent state-of-the-art requires an expensive full SVD in each iteration. In this paper, we show that for many commonly-used nonconvex low-rank regularizers, the singular values obtained from the proximal operator can be automatically threshold. This allows the proximal operator to be efficiently approximated by the power method. We then develop a fast proximal algorithm and its accelerated variant with inexact proximal step. It can be guaranteed that the squared distance between consecutive iterates converges at a rate of $O(1/T)$O(1/T), where $T$T is the number of iterations. Furthermore, we show the proposed algorithm can be parallelized, and the resultant algorithm achieves nearly linear speedup w.r.t. the number of threads. Extensive experiments are performed on matrix completion and robust principal component analysis. Significant speedup over the state-of-the-art is observed. [ABSTRACT FROM AUTHOR]

Programming hybrid CPU-GPU clusters is hard. This paper addresses this difficulty and presents the design and runtime implementation of Unicorn—a parallel programming model for hybrid CPU-GPU clusters. In particular, this paper proves that efficient distributed shared memory style programing is possible and its simplicity can be retained across CPUs and GPUs in a cluster, minus the frustration of dealing with race conditions. Further, this can be done with a unified abstraction, avoiding much of the complication of dealing with hybrid architectures. This is achieved with the help of transactional semantics (on shared global address spaces), deferred bulk data synchronization, workload pipelining and various communication and computation scheduling optimizations. We describe the said abstraction, our computation and communication scheduling system and report its performance on a few benchmarks like Matrix Multiplication, LU Decomposition and 2D FFT. We find that parallelization of coarse-grained applications like matrix multiplication or 2D FFT using our system requires only about 30 lines of C code to set up the runtime. The rest of the application code is regular single CPU/GPU implementation. This indicates the ease of extending parallel code to a distributed environment. The execution is efficient as well. When multiplying two square matrices of size $65,536\times 65,536$ , Unicornachieves a peak performance of 7.88 TFlop/s when run over a cluster of 14 nodes with each node equipped with two Tesla M2070 GPUs and two 6-core Intel Xeon 2.67 GHz CPUs, connected over a 32 Gbps Infiniband network. In this paper, we also demonstrate that the Unicorn programming model can be efficiently used to implement high level abstractions like MapReduce. We use such an extension to implement PageRank and report its performance. For a sample web of 500 million web pages, our implementation completes a page rank iteration in about 18 seconds (on average) on a 14-node cluster. [ABSTRACT FROM AUTHOR]

To cope with uncertainties and variations that emanate from hardware and/or application characteristics, dynamic power management (DPM) frameworks must be able to learn about the system inputs and environmental variations, and adjust the power management policy on the fly. In this paper, an online adaptive DPM technique is presented based on the model-free reinforcement learning (RL) method, which requires no prior knowledge of the state transition probability function and the reward function. In particular, this paper employs the temporal difference (TD) learning method for semi-Markov decision process (SMDP) as the model-free RL technique since the TD method can accelerate convergence and alleviate the reliance on the Markovian property of the power-managed system. In addition, a novel workload predictor based on an online Bayesian classifier is presented to provide effective estimation of the workload characteristics for the RL algorithm. Several improvements are proposed to manage the size of the action space for the learning algorithm, enhance its convergence speed, and dynamically change the action set associated with each system state. In the proposed DPM framework, power-latency tradeoffs of the power-managed system can be precisely controlled based on a user-defined parameter. Extensive experiments on hard disk drives and wireless network cards show that the maximum power saving without sacrificing any latency is 18.6 percent compared to a reference expert-based approach. Alternatively, the maximum latency saving without any power dissipation increase is 73.0 percent compared to the existing best-of-breed DPM techniques. [ABSTRACT FROM PUBLISHER]

This paper addresses the problem of balancing the on-chip packet latencies in a chip multi-processor (CMP), which is simultaneously executing multiple applications. Specifically, this paper presents a balanced application-to-core mapping algorithm that aims to minimize the maximum on-chip packet latency of all running applications. The paper starts by formulating the balanced mapping problem for CMPs and proving its NP-completeness. Next it presents an efficient heuristic algorithm for solving the aforesaid problem, which utilizes the characteristics of on-chip cache and memory accesses in CMPs and takes into account the workload variations among applications. Simulation results on PARSEC benchmark suite show that the proposed algorithm lowers the maximum average packet latency of all applications by 11 percent while cutting the standard deviation of on-chip packet latencies by 99 percent. This is achieved by very little overhead in terms of the overall packet latency and power consumption averaged over all packets. [ABSTRACT FROM AUTHOR]

The papers in this special section focus on software, hardware, and computer applications for neuromorphic computing. Inspired by biological neural systems, neuromorphic computing has drawn much attention for its great potential of achieving machine intelligence at extremely low energy dissipation. Bio-inspired computing models have been investigated for information encoding, sparse representation, event driven communication/computation, and online learning. This new computing paradigm triggered a recent wave of innovations in software and hardware architecture and emerging device technology, which consequently enabled many novel applications. This is an exemplar of research area where the application, computing model, architecture and circuit level design are tightly coupled to deliver unprecedented functionality and energy efficiency. [ABSTRACT FROM AUTHOR]

The Android operating system (OS) is a sophisticated man-made system and is the dominant OS in the current smartphone market. Due to the accumulation of errors in the system internal state and the incremental consumption of resources, such as the Dalvik heap memory of software applications and the physical memory, software aging is observed frequently and recognized as a chronic problem of Android smartphones. To mitigate this problem, we propose a two-level software rejuvenation, with the two levels referring to software applications and the OS, in this paper. Based on this strategy, a Markov regenerative process model is constructed to evaluate the steady-state availability and to optimize the time required to trigger rejuvenation for Android smartphones. The parameters of the model, such as the degradation rate and failure rate of software applications and the Android OS, are obtained via our testing platform. Experiments on two real Android applications show that the availability of an Android smartphone increases by 10.81% and 10.18% for the two subjects in our experiments, respectively. An empirical study comparing our two-level strategy with one-level strategies (single application-level and system-level rejuvenation) further verifies the effectiveness of our approach. [ABSTRACT FROM AUTHOR]

In real-time systems, the application’s behavior has to be predictable at compile-time to guarantee timing constraints. However, modern streaming applications which exhibit adaptive behavior due to mode switching at run-time, may degrade system predictability due to unknown behavior of the application during mode transitions. Therefore, proper temporal analysis during mode transitions is imperative to preserve system predictability. To this end, in this paper, we initially introduce mode-aware data flow (MADF) which is our new predictable model of computation to efficiently capture the behavior of adaptive streaming applications. Then, as an important part of the operational semantics of MADF, we propose the maximum-overlap offset which is our novel protocol for mode transitions. The main advantage of this transition protocol is that, in contrast to self-timed transition protocols, it avoids timing interference between modes upon mode transitions. As a result, any mode transition can be analyzed independently from the mode transitions that occurred in the past. Based on this transition protocol, we propose a hard real-time analysis as well to guarantee timing constraints by avoiding processor overloading during mode transitions. Therefore, using this protocol, we can derive a lower bound and an upper bound on the earliest starting time of the tasks in the new mode during mode transitions in such a way that hard real-time constraints are respected. [ABSTRACT FROM AUTHOR]

In many applications, top-k query is an important operation to return a set of interesting points in a potentially huge data space. It is analyzed in this paper that the existing algorithms cannot process top- k query on massive data efficiently. This paper proposes a novel table-scan-based T2S algorithm to efficiently compute top-k results on massive data. T2S first constructs the presorted table, whose tuples are arranged in the order of the round-robin retrieval on the sorted lists. T2S maintains only fixed number of tuples to compute results. The early termination checking for T2S is presented in this paper, along with the analysis of scan depth. The selective retrieval is devised to skip the tuples in the presorted table which are not top-k results. The theoretical analysis proves that selective retrieval can reduce the number of the retrieved tuples significantly. The construction and incremental-update/batch-processing methods for the used structures are proposed in this paper. The extensive experimental results, conducted on synthetic and real-life data sets, show that T2S has a significant advantage over the existing algorithms. [ABSTRACT FROM AUTHOR]

With increasing demands on cloud computing, protecting guest virtual machines (VMs) from malicious attackers has become critical to provide secure services. The current cloud security model with software-based virtualization relies on the invulnerability of the software hypervisor and its trustworthy administrator with the root permission. However, compromising the hypervisor with remote attacks or root permission grants the attackers with a full access capability to the memory and context of a guest VM. This paper proposes a HW-based approach to protect guest VMs even under an untrusted hypervisor. With the proposed mechanism, memory isolation is provided by the secure hardware, which is much less vulnerable than the software hypervisor. The proposed mechanism extends the current hardware support for memory virtualization based on nested paging with a small extra hardware cost. The hypervisor can still flexibly allocate physical memory pages to virtual machines for efficient resource management. In addition to the system design for secure virtualization, this paper presents a prototype implementation using system management mode. Although the current system management mode is not intended for security functions and thus limits the performance and complete protection, the prototype implementation proves the feasibility of the proposed design. [ABSTRACT FROM AUTHOR]

Lambda paradigm, also known as Function as a Service (FaaS), is a novel event-driven concept that allows companies to build scalable and reliable enterprise applications in an off-premise computing data-center as a serverless solution. In practice, however, an important goal for the service provider of a Lambda platform is to devise an efficient way to consolidate multiple Lambda functions in a single host. While the majority of existing resource management solutions use only operating-system level metrics (e.g., average utilization of computing and I/O resources) to allocate the available resources among the submitted workloads in a balanced way, a resource allocation schema that is oblivious to the issue of shared-resource contention can result in a significant performance variability and degradation within the entire platform. This paper proposes a predictive controller scheme that dynamically allocates resources in a Lambda platform. This scheme uses a prediction tool to estimate the future rate of every event stream and takes into account the quality of service enforcements requested by the owner of each Lambda function. This is formulated as an optimization problem where a set of cost functions are introduced (i) to reduce the total QoS violation incidents; (ii) to keep the CPU utilization level within an accepted range; and (iii) to avoid the fierce contention among collocated applications for obtaining shared resources. Performance evaluation is carried out by comparing the proposed solution with an enhanced interference-aware version of three well-known heuristics, namely spread, binpack (the two native clustering solutions employed by Docker Swarm) and best-effort resource allocation schema. Experimental results show that the proposed controller improves the overall performance (in terms of reducing the end-to-end response time) by 14.9 percent on average compared to the best result of the other heuristics. The proposed solution also increases the overall CPU utilization by 18 percent on average (for lightweight workloads), while achieves an average 87 percent (maximum 146 percent) improvement in preventing QoS violation incidents. [ABSTRACT FROM AUTHOR]

Reliability to soft errors is an increasingly important issue as technology continues to shrink. In this paper, we show that applications exhibit different reliability characteristics on big, high-performance cores versus small, power-efficient cores, and that there is significant opportunity to improve system reliability through reliability-aware scheduling on heterogeneous multicore processors. We monitor the reliability characteristics of all running applications, and dynamically schedule applications to the different core types in a heterogeneous multicore to maximize system reliability. Reliability-aware scheduling improves reliability by 25.4 percent on average (and up to 60.2 percent) compared to performance-optimized scheduling on a heterogeneous multicore processor with two big cores and two small cores, while degrading performance by 6.3 percent only. We also introduce a novel system-level reliability metric for multiprogram workloads on (heterogeneous) multicores. We provide a trade-off analysis among reliability-, power- and performance-optimized scheduling, and evaluate reliability-aware scheduling under performance constraints and for unprotected L1 caches. In addition, we also extend our scheduling mechanisms to multithreaded programs. The hardware cost in support of our reliability-aware scheduler is limited to 296 bytes per core. [ABSTRACT FROM PUBLISHER]

In this paper, we report about platform and architecture that real-time analysis of big data are possible, and structured IT infrastructure that they are optimally combined. We developed a distributed architecture which the data conversion and the abnormality determination are multi-blocked. Furthermore, by selecting a distributed storage database (DB), we succeeded in constructing IT infrastructure capable of high-speed processing at a large number of manufacturing sites. In the new IT infrastructure, we achieved resource leveling of the application server and improvement of data processing time. It is expected that data file stagnation and delay of DB registration can be resolved. [ABSTRACT FROM AUTHOR]

The data stored in a static random access memory (SRAM) immediately after power-up is of practical interest for some applications, such as SRAM physical unclonable functions. In this paper, measurements of SRAM cell power-up state (i.e., whether the cell is storing 0 or 1 after the power supply is turned on) using an addressable cell array test structure are reported. The test structure provides direct access to individual transistor characteristics of many SRAM cells, which would facilitate the characterization of SRAM power-up behavior. Methods and considerations necessary for reliable and stable power-up state characterization using the test structure will be discussed and demonstrated. [ABSTRACT FROM PUBLISHER]

Rollback is an effective technique to resume the system execution from a recorded intermediate state upon failures, without having to restart the entire system. However, in virtualized environments, rollback of a virtual machine cluster (VMC) produces high network traffic and long service disruption, particularly for a large cluster used for scientific computing, thereby imposing significant overhead both on network and applications. This paper proposes Piccolo, a fast and efficient rollback system, to restore a VMC from snapshot files over data center network. First, we exploit the similarity among VMC snapshots and leverage multicast to deliver the identical pages across VMs placed on disperse hosts, thereby bypassing unnecessary transmission of a large number of pages. Second, we analyze the impact on network traffic of varying VM placements in data center network, formulate the traffic aware placement as an optimization problem, and design a two-tier approximation algorithm that efficiently solves the problem. In addition to presenting Piccolo, we detail its implementation, and evaluate it by a set of experiments. The results show that Piccolo could achieve a significant reduction in terms of total sent data, network traffic and rollback latency compared to the existing generic techniques. [ABSTRACT FROM AUTHOR]

In this paper, we study the set reconciliation problem, in which each member of a node pair has a set of objects and seeks to deliver its unique objects to the other member. How could each node compute the set difference, however, is challenging in the set reconciliation problem. To address such an issue, we propose a lightweight but efficient method that only requires the pair of nodes to represent objects using a counting Bloom filter (CBF) of size $(O(d))$ and exchange with each other, where $(d)$ denotes the total size of the set differences. A receiving node then subtracts the received CBF from its local one via minus operation proposed in this paper. The resultant CBF can approximately represent the union of the set differences and thus the set difference to each node can be identified after querying the resultant CBF. In this paper, we propose a novel estimator through which each node can accurately estimate not only the value of $(d)$ but also the size of the set difference to each node. Such an estimation result can be used to optimize the parameter setting of the CBF to achieve less false positives and false negatives. Comprehensive analysis and evaluation demonstrates that our method is more efficient than prior BF-based methods in terms of achieving the same accuracy with less communication cost. Moreover, our reconciliating method needs no prior context logs and it is very useful in networking and distributed applications. [ABSTRACT FROM AUTHOR]

In this paper, a low-complexity adaptive edge-enhanced algorithm is proposed for the implementation of 2-D image scaling applications. The proposed novel algorithm consists of a linear space-variant edge detector, a low complexity sharpening spatial filter, and a simplified bilinear interpolation. The edge detector is designed to discover the image edges by a low-cost edge-catching technique. The sharpening spatial filter is added as a prefilter to reduce the blurring effect produced by the bilinear interpolation. Furthermore, an adaptive technology is used to enhance the effect of the edge detector by adaptively selecting the input pixels of the bilinear interpolation. In addition, an algebraic manipulation and a hardware sharing techniques are used to simplify bilinear interpolation, which efficiently reduces the computing resources and silicon area in very large scale integration (VLSI) circuits. By adding eight 8-bit registers as a register bank, this design can process streaming data directly and requires only a one-line-buffer memory. The VLSI architecture of this paper contains 6.67-K gate counts and achieves about 280-MHz processing rate by using the TSMC 0.13-um CMOS process. Compared with previous low-complexity techniques, this paper performs with better quality, higher performance, less memory requirements, and lower hardware cost than other image scaling methods. [ABSTRACT FROM PUBLISHER]

As a result of their involvement in several research and development and innovation (R&D&I) projects developed in various companies, the authors acquired knowledge of two basic concepts: quality and innovation. The application of these concepts in the teaching–learning process can help teachers to incorporate informal changes in the curriculum to provide for continuous improvement. This paper presents a simple methodology to apply quality and innovation concepts and tools in the daily teaching activity. In order to show the ease of use, as well as the efficiency and effectiveness of this methodology, the experience of applying it in an engineering course over nine years, the actions undertaken, the results obtained, and the lessons learned are described here. The methodology incorporates reflection, innovation, and decision taking based on objective data. Contact with the companies involved has enabled the teachers to learn continuously and to enhance their motivation when working with students. The methodology applied and the example described in this paper can be a starting point for other teachers interested in improving the quality of their teaching. [ABSTRACT FROM AUTHOR]

The identification of services during the execution of service-based applications to replace services in them that are no longer available and/or fail to satisfy certain requirements is an important issue. In this paper, we present a framework to support runtime service discovery. This framework can execute service discovery queries in pull and push mode. In pull mode, it executes queries when a need for finding a replacement service arises. In push mode, queries are subscribed to the framework to be executed proactively and, in parallel with the operation of the application, to identify adequate services that could be used if the need for replacing a service arises. Hence, the proactive (push) mode of query execution makes it more likely to avoid interruptions in the operation of service-based applications when a service in them needs to be replaced at runtime. In both modes of query execution, the identification of services relies on distance-based matching of structural, behavioral, quality, and contextual characteristics of services and applications. A prototype implementation of the framework has been developed and an evaluation was carried out to assess the performance of the framework. This evaluation has shown positive results, which are discussed in the paper. [ABSTRACT FROM PUBLISHER]

In this paper, we explore a new form of dissemination that arises in distributed, mission-critical applications called Flash Dissemination. This involves the rapid dissemination of rich information to a large number of recipients in a very short period of time. A key characteristic of Flash Dissemination is its unpredictability (e.g., natural hazards), but when invoked it must harness all possible resources to ensure timely delivery of information. Additionally, it must scale to a large number of recipients and perform efficiently in highly heterogeneous (data, network) and failure prone environments. We investigate a peer-based approach based on the simple principle of transferring dissemination load to information receivers using foundations from broadcast networks, gossip theory, and random networks. Gossip-based protocols are well known for being stateless, scalable, and fault-tolerant; however, their performance degrades as content size increases, because of the propagation of redundant gossip messages. In this paper, we propose Concurrent Random Expanding Walkers (CREW), a smart gossip protocol designed to maximize the speed of dissemination by transmitting data only as needed, and by exploiting both intra- and internode concurrency. CREW is designed to support both content and network heterogeneity and deal with transmission failures without sacrificing dissemination speed. We implemented CREW on top of a scalable middleware environment that allows for deployment across several platforms and developed optimizations without compromising on the stateless nature of CREW. We evaluated CREW empirically and compared it to optimized implementations of popular gossip and peer-based systems. Our experiments show that CREW significantly outperforms both traditional gossip and current large content dissemination systems while sustaining its performance in the presence of network errors. [ABSTRACT FROM AUTHOR]

This paper describes developments in high-altitude electromagnetic pulse (HEMP) test methods. More specifically, we describe the most commonly used HEMP test methods and associated measurement systems used to assess systems and equipment to the early-time (E1) effects of HEMP. Short case studies of the application of these test methods are included. Finally, computational electromagnetics (CEM) is a subject that has advanced significantly in recent years. This paper summarizes developments in this area and the supporting role that CEM has when undertaking assessments of the impact of HEMP on systems. [ABSTRACT FROM AUTHOR]