Xidian University

As a convex relaxation of the low rank matrix factorization problem, the nuclear norm minimization has been attracting significant research interest in recent years. The standard nuclear norm minimization regularizes each singular value equally to pursue the convexity of the objective function. However, this greatly restricts its capability and flexibility in dealing with many practical problems (e.g., denoising), where the singular values have clear physical meanings and should be treated differently. In this paper we study the weighted nuclear norm minimization (WNNM) problem, where the singular values are assigned different weights. The solutions of the WNNM problem are analyzed under different weighting conditions. We then apply the proposed WNNM algorithm to image denoising by exploiting the image nonlocal self-similarity. Experimental results clearly show that the proposed WNNM algorithm outperforms many state-of-the-art denoising algorithms such as BM3D in terms of both quantitative measure and visual perception quality.

As a recently proposed technique, sparse representation based classification (SRC) has been widely used for face recognition (FR). SRC first codes a testing sample as a sparse linear combination of all the training samples, and then classifies the testing sample by evaluating which class leads to the minimum representation error. While the importance of sparsity is much emphasized in SRC and many related works, the use of collaborative representation (CR) in SRC is ignored by most literature. However, is it really the l <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> -norm sparsity that improves the FR accuracy? This paper devotes to analyze the working mechanism of SRC, and indicates that it is the CR but not the l1-norm sparsity that makes SRC powerful for face classification. Consequently, we propose a very simple yet much more efficient face classification scheme, namely CR based classification with regularized least square (CRC_RLS). The extensive experiments clearly show that CRC_RLS has very competitive classification results, while it has significantly less complexity than SRC.

In this paper, we present {GraRep}, a novel model for learning vertex representations of weighted graphs. This model learns low dimensional vectors to represent vertices appearing in a graph and, unlike existing work, integrates global structural information of the graph into the learning process. We also formally analyze the connections between our work and several previous research efforts, including the DeepWalk model of Perozzi et al. as well as the skip-gram model with negative sampling of Mikolov et al.

Sparse representation models code an image patch as a linear combination of a few atoms chosen out from an over-complete dictionary, and they have shown promising results in various image restoration applications. However, due to the degradation of the observed image (e.g., noisy, blurred, and/or down-sampled), the sparse representations by conventional models may not be accurate enough for a faithful reconstruction of the original image. To improve the performance of sparse representation-based image restoration, in this paper the concept of sparse coding noise is introduced, and the goal of image restoration turns to how to suppress the sparse coding noise. To this end, we exploit the image nonlocal self-similarity to obtain good estimates of the sparse coding coefficients of the original image, and then centralize the sparse coding coefficients of the observed image to those estimates. The so-called nonlocally centralized sparse representation (NCSR) model is as simple as the standard sparse representation model, while our extensive experiments on various types of image restoration problems, including denoising, deblurring and super-resolution, validate the generality and state-of-the-art performance of the proposed NCSR algorithm.

Object detection is one of the most important and challenging branches of computer vision, which has been widely applied in people’s life, such as monitoring security, autonomous driving and so on, with the purpose of locating instances of semantic objects of a certain class. With the rapid development of deep learning algorithms for detection tasks, the performance of object detectors has been greatly improved. In order to understand the main development status of object detection pipeline thoroughly and deeply, in this survey, we analyze the methods of existing typical detection models and describe the benchmark datasets at first. Afterwards and primarily, we provide a comprehensive overview of a variety of object detection methods in a systematic manner, covering the one-stage and two-stage detectors. Moreover, we list the traditional and new applications. Some representative branches of object detection are analyzed as well. Finally, we discuss the architecture of exploiting these object detection methods to build an effective and efficient system and point out a set of development trends to better follow the state-of-the-art algorithms and further research.

Space-air-ground integrated network (SAGIN), as an integration of satellite systems, aerial networks, and terrestrial communications, has been becoming an emerging architecture and attracted intensive research interest during the past years. Besides bringing significant benefits for various practical services and applications, SAGIN is also facing many unprecedented challenges due to its specific characteristics, such as heterogeneity, self-organization, and time-variability. Compared to traditional ground or satellite networks, SAGIN is affected by the limited and unbalanced network resources in all three network segments, so that it is difficult to obtain the best performances for traffic delivery. Therefore, the system integration, protocol optimization, resource management, and allocation in SAGIN is of great significance. To the best of our knowledge, we are the first to present the state-of-the-art of the SAGIN since existing survey papers focused on either only one single network segment in space or air, or the integration of space-ground, neglecting the integration of all the three network segments. In light of this, we present in this paper a comprehensive review of recent research works concerning SAGIN from network design and resource allocation to performance analysis and optimization. After discussing several existing network architectures, we also point out some technology challenges and future directions.

As a powerful statistical image modeling technique, sparse representation has been successfully used in various image restoration applications. The success of sparse representation owes to the development of the l(1)-norm optimization techniques and the fact that natural images are intrinsically sparse in some domains. The image restoration quality largely depends on whether the employed sparse domain can represent well the underlying image. Considering that the contents can vary significantly across different images or different patches in a single image, we propose to learn various sets of bases from a precollected dataset of example image patches, and then, for a given patch to be processed, one set of bases are adaptively selected to characterize the local sparse domain. We further introduce two adaptive regularization terms into the sparse representation framework. First, a set of autoregressive (AR) models are learned from the dataset of example image patches. The best fitted AR models to a given patch are adaptively selected to regularize the image local structures. Second, the image nonlocal self-similarity is introduced as another regularization term. In addition, the sparsity regularization parameter is adaptively estimated for better image restoration performance. Extensive experiments on image deblurring and super-resolution validate that by using adaptive sparse domain selection and adaptive regularization, the proposed method achieves much better results than many state-of-the-art algorithms in terms of both PSNR and visual perception.

Recently, fabricated two dimensional (2D) phosphorene crystal structures have demonstrated great potential in applications of electronics. Mechanical strain was demonstrated to be able to significantly modify the electronic properties of phosphorene and few-layer black phosphorus. In this work, we employed first principles density functional theory calculations to explore the mechanical properties of phosphorene, including ideal tensile strength and critical strain. It was found that a monolayer phosphorene can sustain tensile strain up to 27% and 30% in the zigzag and armchair directions, respectively. This enormous strain limit of phosphorene results from its unique puckered crystal structure. We found that the tensile strain applied in the armchair direction stretches the pucker of phosphorene, rather than significantly extending the P-P bond lengths. The compromised dihedral angles dramatically reduce the required strain energy. Compared to other 2D materials, such as graphene, phosphorene demonstrates superior flexibility with an order of magnitude smaller Young's modulus. This is especially useful in practical large-magnitude-strain engineering. Furthermore, the anisotropic nature of phosphorene was also explored. We derived a general model to calculate the Young's modulus along different directions for a 2D system.

In this paper, we propose a novel model for learning graph representations, which generates a low-dimensional vector representation for each vertex by capturing the graph structural information. Different from other previous research efforts, we adopt a random surfing model to capture graph structural information directly, instead of using the sampling-based method for generating linear sequences proposed by Perozzi et al. (2014). The advantages of our approach will be illustrated from both theorical and empirical perspectives. We also give a new perspective for the matrix factorization method proposed by Levy and Goldberg (2014), in which the pointwise mutual information (PMI) matrix is considered as an analytical solution to the objective function of the skip-gram model with negative sampling proposed by Mikolov et al. (2013). Unlike their approach which involves the use of the SVD for finding the low-dimensitonal projections from the PMI matrix, however, the stacked denoising autoencoder is introduced in our model to extract complex features and model non-linearities. To demonstrate the effectiveness of our model, we conduct experiments on clustering and visualization tasks, employing the learned vertex representations as features. Empirical results on datasets of varying sizes show that our model outperforms other stat-of-the-art models in such tasks.

Abstract: As a powerful statistical image modeling technique, sparse representation has been successfully used in various image restoration applications. The success of sparse representation owes to the development of l1-norm optimization techniques, and the fact that natural images are intrinsically sparse in some domain. The image restoration quality largely depends on whether the employed sparse domain can represent well the underlying image. Considering that the contents can vary significantly across different images or different patches in a single image, we propose to learn various sets of bases from a pre-collected dataset of example image patches, and then for a given patch to be processed, one set of bases are adaptively selected to characterize the local sparse domain. We further introduce two adaptive regularization terms into the sparse representation framework. First, a set of autoregressive (AR) models are learned from the dataset of example image patches. The best fitted AR models to a given patch are adaptively selected to regularize the image local structures. Second, the image non-local self-similarity is introduced as another regularization term. In addition, the sparsity regularization parameter is adaptively estimated for better image restoration performance. Extensive experiments on image deblurring and super-resolution validate that by using adaptive sparse domain selection and adaptive regularization, the proposed method achieves much better results than many state-of-the-art algorithms in terms of both PSNR and visual perception.

The incorporation of dynamic voltage scaling technology into computation offloading offers more flexibilities for mobile edge computing. In this paper, we investigate partial computation offloading by jointly optimizing the computational speed of smart mobile device (SMD), transmit power of SMD, and offloading ratio with two system design objectives: energy consumption of SMD minimization (ECM) and latency of application execution minimization (LM). Considering the case that the SMD is served by a single cloud server, we formulate both the ECM problem and the LM problem as nonconvex problems. To tackle the ECM problem, we recast it as a convex one with the variable substitution technique and obtain its optimal solution. To address the nonconvex and nonsmooth LM problem, we propose a locally optimal algorithm with the univariate search technique. Furthermore, we extend the scenario to a multiple cloud servers system, where the SMD could offload its computation to a set of cloud servers. In this scenario, we obtain the optimal computation distribution among cloud servers in closed form for the ECM and LM problems. Finally, extensive simulations demonstrate that our proposed algorithms can significantly reduce the energy consumption and shorten the latency with respect to the existing offloading schemes.

Medical imaging can assess the tumor and its environment in their entirety, which makes it suitable for monitoring the temporal and spatial characteristics of the tumor. Progress in computational methods, especially in artificial intelligence for medical image process and analysis, has converted these images into quantitative and minable data associated with clinical events in oncology management. This concept was first described as radiomics in 2012. Since then, computer scientists, radiologists, and oncologists have gravitated towards this new tool and exploited advanced methodologies to mine the information behind medical images. On the basis of a great quantity of radiographic images and novel computational technologies, researchers developed and validated radiomic models that may improve the accuracy of diagnoses and therapy response assessments. Here, we review the recent methodological developments in radiomics, including data acquisition, tumor segmentation, feature extraction, and modelling, as well as the rapidly developing deep learning technology. Moreover, we outline the main applications of radiomics in diagnosis, treatment planning and evaluations in the field of oncology with the aim of developing quantitative and personalized medicine. Finally, we discuss the challenges in the field of radiomics and the scope and clinical applicability of these methods.

In recent years, single image super-resolution (SISR) methods using deep convolution neural network (CNN) have achieved impressive results. Thanks to the powerful representation capabilities of the deep networks, numerous previous ways can learn the complex non-linear mapping between low-resolution (LR) image patches and their high-resolution (HR) versions. However, excessive convolutions will limit the application of super-resolution technology in low computing power devices. Besides, super-resolution of any arbitrary scale factor is a critical issue in practical applications, which has not been well solved in the previous approaches. To address these issues, we propose a lightweight information multi-distillation network (IMDN) by constructing the cascaded information multi-distillation blocks (IMDB), which contains distillation and selective fusion parts. Specifically, the distillation module extracts hierarchical features step-by-step, and fusion module aggregates them according to the importance of candidate features, which is evaluated by the proposed contrast-aware channel attention mechanism. To process real images with any sizes, we develop an adaptive cropping strategy (ACS) to super-resolve block-wise image patches using the same well-trained model. Extensive experiments suggest that the proposed method performs favorably against the state-of-the-art SR algorithms in term of visual quality, memory footprint, and inference time. Code is available at \urlhttps://github.com/Zheng222/IMDN.

Recently, deep convolutional neural networks (CNNs) have been demonstrated remarkable progress on single image super-resolution. However, as the depth and width of the networks increase, CNN-based super-resolution methods have been faced with the challenges of computational complexity and memory consumption in practice. In order to solve the above questions, we propose a deep but compact convolutional network to directly reconstruct the high resolution image from the original low resolution image. In general, the proposed model consists of three parts, which are feature extraction block, stacked information distillation blocks and reconstruction block respectively. By combining an enhancement unit with a compression unit into a distillation block, the local long and short-path features can be effectively extracted. Specifically, the proposed enhancement unit mixes together two different types of features and the compression unit distills more useful information for the sequential blocks. In addition, the proposed network has the advantage of fast execution due to the comparatively few numbers of filters per layer and the use of group convolution. Experimental results demonstrate that the proposed method is superior to the state-of-the-art methods, especially in terms of time performance. Code is available at https://github.com/Zheng222/IDN-Caffe.

Macrophages play an important role in cancer development and metastasis. Proinflammatory M1 macrophages can phagocytose tumor cells, while anti-inflammatory M2 macrophages such as tumor-associated macrophages (TAMs) promote tumor growth and invasion. Modulating the tumor immune microenvironment through engineering macrophages is efficacious in tumor therapy. M1 macrophages target cancerous cells and, therefore, can be used as drug carriers for tumor therapy. Herein, the strategies to engineer macrophages for cancer immunotherapy, such as inhibition of macrophage recruitment, depletion of TAMs, reprograming of TAMs, and blocking of the CD47-SIRPα pathway, are discussed. Further, the recent advances in drug delivery using M1 macrophages, macrophage-derived exosomes, and macrophage-membrane-coated nanoparticles are elaborated. Overall, there is still significant room for development in macrophage-mediated immune modulation and macrophage-mediated drug delivery, which will further enhance current tumor therapies against various malignant solid tumors, including drug-resistant tumors and metastatic tumors.

Recently fabricated two-dimensional phosphorene crystal structures have demonstrated great potential in applications of electronics. In this paper, strain effect on the electronic band structure of phosphorene was studied using first-principles methods including density functional theory (DFT) and hybrid functionals. It was found that phosphorene can withstand a tensile stress and strain up to 10 N/m and 30%, respectively. The band gap of phosphorene experiences a direct-indirect-direct transition when axial strain is applied. A moderate \ensuremath{-}2% compression in the zigzag direction can trigger this gap transition. With sufficient expansion (+11.3%) or compression (\ensuremath{-}10.2% strains), the gap can be tuned from indirect to direct again. Five strain zones with distinct electronic band structure were identified, and the critical strains for the zone boundaries were determined. Although the DFT method is known to underestimate band gap of semiconductors, it was proven to correctly predict the strain effect on the electronic properties with validation from a hybrid functional method in this work. The origin of the gap transition was revealed, and a general mechanism was developed to explain energy shifts with strain according to the bond nature of near-band-edge electronic orbitals. Effective masses of carriers in the armchair direction are an order of magnitude smaller than that of the zigzag axis, indicating that the armchair direction is favored for carrier transport. In addition, the effective masses can be dramatically tuned by strain, in which its sharp jump/drop occurs at the zone boundaries of the direct-indirect gap transition.

Sparse representation based classification has led to interesting image recognition results, while the dictionary used for sparse coding plays a key role in it. This paper presents a novel dictionary learning (DL) method to improve the pattern classification performance. Based on the Fisher discrimination criterion, a structured dictionary, whose dictionary atoms have correspondence to the class labels, is learned so that the reconstruction error after sparse coding can be used for pattern classification. Meanwhile, the Fisher discrimination criterion is imposed on the coding coefficients so that they have small within-class scatter but big between-class scatter. A new classification scheme associated with the proposed Fisher discrimination DL (FDDL) method is then presented by using both the discriminative information in the reconstruction error and sparse coding coefficients. The proposed FDDL is extensively evaluated on benchmark image databases in comparison with existing sparse representation and DL based classification methods.

Mobile-edge computing (MEC) and wireless power transfer are two promising techniques to enhance the computation capability and to prolong the operational time of low-power wireless devices that are ubiquitous in Internet of Things. However, the computation performance and the harvested energy are significantly impacted by the severe propagation loss. In order to address this issue, an unmanned aerial vehicle (UAV)-enabled MEC wireless-powered system is studied in this paper. The computation rate maximization problems in a UAV-enabled MEC wireless powered system are investigated under both partial and binary computation offloading modes, subject to the energy-harvesting causal constraint and the UAV's speed constraint. These problems are non-convex and challenging to solve. A two-stage algorithm and a three-stage alternative algorithm are, respectively, proposed for solving the formulated problems. The closed-form expressions for the optimal central processing unit frequencies, user offloading time, and user transmit power are derived. The optimal selection scheme on whether users choose to locally compute or offload computation tasks is proposed for the binary computation offloading mode. Simulation results show that our proposed resource allocation schemes outperform other benchmark schemes. The results also demonstrate that the proposed schemes converge fast and have low computational complexity.

Internet of Things (IoT) computing offloading is a challenging issue, especially in remote areas where common edge/cloud infrastructure is unavailable. In this paper, we present a space-air-ground integrated network (SAGIN) edge/cloud computing architecture for offloading the computation-intensive applications considering remote energy and computation constraints, where flying unmanned aerial vehicles (UAVs) provide near-user edge computing and satellites provide access to the cloud computing. First, for UAV edge servers, we propose a joint resource allocation and task scheduling approach to efficiently allocate the computing resources to virtual machines (VMs) and schedule the offloaded tasks. Second, we investigate the computing offloading problem in SAGIN and propose a learning-based approach to learn the optimal offloading policy from the dynamic SAGIN environments. Specifically, we formulate the offloading decision making as a Markov decision process where the system state considers the network dynamics. To cope with the system dynamics and complexity, we propose a deep reinforcement learning-based computing offloading approach to learn the optimal offloading policy on-the-fly, where we adopt the policy gradient method to handle the large action space and actor-critic method to accelerate the learning process. Simulation results show that the proposed edge VM allocation and task scheduling approach can achieve near-optimal performance with very low complexity and the proposed learning-based computing offloading algorithm not only converges fast but also achieves a lower total cost compared with other offloading approaches.

BACKGROUND: High-throughput transcriptome sequencing (RNA-seq) technology promises to discover novel protein-coding and non-coding transcripts, particularly the identification of long non-coding RNAs (lncRNAs) from de novo sequencing data. This requires tools that are not restricted by prior gene annotations, genomic sequences and high-quality sequencing. RESULTS: We present an alignment-free tool called PLEK (predictor of long non-coding RNAs and messenger RNAs based on an improved k-mer scheme), which uses a computational pipeline based on an improved k-mer scheme and a support vector machine (SVM) algorithm to distinguish lncRNAs from messenger RNAs (mRNAs), in the absence of genomic sequences or annotations. The performance of PLEK was evaluated on well-annotated mRNA and lncRNA transcripts. 10-fold cross-validation tests on human RefSeq mRNAs and GENCODE lncRNAs indicated that our tool could achieve accuracy of up to 95.6%. We demonstrated the utility of PLEK on transcripts from other vertebrates using the model built from human datasets. PLEK attained >90% accuracy on most of these datasets. PLEK also performed well using a simulated dataset and two real de novo assembled transcriptome datasets (sequenced by PacBio and 454 platforms) with relatively high indel sequencing errors. In addition, PLEK is approximately eightfold faster than a newly developed alignment-free tool, named Coding-Non-Coding Index (CNCI), and 244 times faster than the most popular alignment-based tool, Coding Potential Calculator (CPC), in a single-threading running manner. CONCLUSIONS: PLEK is an efficient alignment-free computational tool to distinguish lncRNAs from mRNAs in RNA-seq transcriptomes of species lacking reference genomes. PLEK is especially suitable for PacBio or 454 sequencing data and large-scale transcriptome data. Its open-source software can be freely downloaded from https://sourceforge.net/projects/plek/files/.