A5100359961- Face and Expression Recognition
- Face recognition and analysis
- Biometric Identification and Security
- Advanced Image and Video Retrieval Techniques
- Sparse and Compressive Sensing Techniques
- Image Retrieval and Classification Techniques
- Video Surveillance and Tracking Methods
- Anomaly Detection Techniques and Applications
- Remote-Sensing Image Classification
- Domain Adaptation and Few-Shot Learning
- Advanced Image Processing Techniques
- Speech and Audio Processing
- Image and Signal Denoising Methods
- Text and Document Classification Technologies
- Advanced Neural Network Applications
- Image Processing Techniques and Applications
- Neural Networks and Applications
- Advanced Image Fusion Techniques
- Machine Learning in Bioinformatics
- Ubiquitin and proteasome pathways
- Protein Degradation and Inhibitors
- Endoplasmic Reticulum Stress and Disease
- Gene expression and cancer classification
- Cell death mechanisms and regulation
- Phagocytosis and Immune Regulation
Harbin Institute of Technology
2016-2025
Shenzhen Institute of Information Technology
2007-2025
Peng Cheng Laboratory
2019-2025
South China University of Technology
2009-2025
Hunan University
2025
Liaocheng University
2021-2025
Nanjing Medical University
2021-2025
China Pharmaceutical University
2025
Nanjing University of Science and Technology
2003-2024
Cloud Computing Center
2015-2024
Sparse representation has attracted much attention from researchers in fields of signal processing, image computer vision and pattern recognition. also a good reputation both theoretical research practical applications. Many different algorithms have been proposed for sparse representation. The main purpose this article is to provide comprehensive study an updated review on supply guidance researchers. taxonomy methods can be studied various viewpoints. For example, terms norm minimizations...
In domain adaptation, maximum mean discrepancy (MMD) has been widely adopted as a metric between the distributions of source and target domains. However, existing MMD-based adaptation methods generally ignore changes class prior distributions, i.e., weight bias across This remains an open problem but ubiquitous for which can be caused by in sample selection criteria application scenarios. We show that MMD cannot account results degraded performance. To address this issue, weighted model is...
Logs are widely used by large and complex software-intensive systems for troubleshooting. There have been a lot of studies on log-based anomaly detection. To detect the anomalies, existing methods mainly construct detection model using log event data extracted from historical logs. However, we find that do not work well in practice. These close-world assumption, which assumes is stable over time set distinct events known. our empirical study shows practice, often contains previously unseen...
Recently, regression analysis has become a popular tool for face recognition. Most existing methods use the one-dimensional, pixel-based error model, which characterizes representation individually, pixel by pixel, and thus neglects two-dimensional structure of image. We observe that occlusion illumination changes generally lead, approximately, to low-rank In order make this structural information, paper presents image-matrix-based namely, nuclear norm based matrix (NMR), classification. NMR...
Linear discriminant analysis (LDA) is a very popular supervised feature extraction method and has been extended to different variants. However, classical LDA the following problems: 1) The obtained projection does not have good interpretability for features; 2) sensitive noise; 3) selection of number directions. In this paper, novel called robust sparse linear (RSLDA) proposed solve above problems. Specifically, RSLDA adaptively selects most discriminative features by introducing...
In this paper, we address the problem of unsupervised domain transfer learning in which no labels are available target domain. We use a transformation matrix to both source and data common subspace, where each sample can be represented by combination samples such that from different domains well interlaced. way, discrepancy is reduced. By imposing joint low-rank sparse constraints on reconstruction coefficient matrix, global local structures preserved. To enlarge margins between classes as...
In this paper, we propose a general framework for incomplete multiview clustering. The proposed method is the first work that exploits graph learning and spectral clustering techniques to learn common representation First, owing good performance of low-rank in discovering intrinsic subspace structure data, adopt it adaptively construct each view. Second, constraint used achieve low-dimensional view based on Third, further introduce co-regularization term samples all views, then use k -means...
In this brief, multilinear sparse principal component analysis (MSPCA) is proposed for feature extraction from the tensor data. MSPCA can be viewed as a further extension of classical (PCA), PCA (SPCA) and recently (MPCA). The key operation to rewrite MPCA into regression forms relax it regression. Differing MPCA, inherits sparsity SPCA iteratively learns series projections that capture most variation Each nonzero element in selected important variables/factors using elastic net. Extensive...
Video and images acquired by a visual system are seriously degraded under hazy foggy weather, which will affect the detection, tracking, recognition of targets. Thus, restoring true scene from such video or image is significance. The main goal this paper was to summarize current defogging algorithms. We first presented review detection classification method image. Then, we summarized existing algorithms, including restoration contrast enhancement fusion-based also objective quality...
Multi-view clustering aims to partition data collected from diverse sources based on the assumption that all views are complete. However, such prior is hardly satisfied in many real-world applications, resulting incomplete multi-view learning problem. The existing attempts this problem still have following limitations: 1) underlying semantic information of missing commonly ignored; 2) local structure not well explored; 3) importance different effectively evaluated. To address these issues,...
Deep convolutional neural networks (CNNs) have been popularly adopted in image super-resolution (SR). However, deep CNNs for SR often suffer from the instability of training, resulting poor performance. Gathering complementary contextual information can effectively overcome problem. Along this line, we propose a coarse-to-fine CNN (CFSRCNN) to recover high-resolution (HR) its low-resolution version. The proposed CFSRCNN consists stack feature extraction blocks (FEBs), an enhancement block...
Sparse representation has shown an attractive performance in a number of applications. However, the available sparse methods still suffer from some problems, and it is necessary to design more efficient methods. Particularly, computationally inexpensive, easily solvable, robust method significant task. In this paper, we explore issue designing simple, robust, powerfully for image classification. The contributions paper are as follows. First, novel discriminative proposed its noticeable...
In this paper, a novel deep network is proposed for multi-focus image fusion, named Deep Regression Pair Learning (DRPL). contrast to existing fusion methods which divide the input into small patches and apply classifier judge whether patch in focus or not, DRPL directly converts whole binary mask without any operation, subsequently tackling difficulty of blur level estimation around focused/defocused boundary. Simultaneously, pair learning strategy, takes complementary source images as...
Recently Transformer has been largely explored in tracking and shown state-of-the-art (SOTA) performance. However, existing efforts mainly focus on fusing enhancing features generated by convolutional neural networks (CNNs). The potential of representation learning remains under-explored. In this paper, we aim to further unleash the power proposing a simple yet efficient fully-attentional tracker, dubbed SwinTrack, within classic Siamese framework. particular, both feature fusion SwinTrack...
In real-world applications, it is often that the collected multi-view data are incomplete, i.e., some views of samples absent. Existing clustering methods for incomplete all focus on obtaining a common representation or graph from available but neglect hidden information missing and imbalance different views. To solve these problems, novel method, called adaptive completion based (AGC_IMC), proposed in this paper. Specifically, AGC_IMC develops joint framework consensus learning, which...
Conventional multi-view clustering seeks to partition data into respective groups based on the assumption that all views are fully observed. However, in practical applications, such as disease diagnosis, multimedia analysis, and recommendation system, it is common observe not of samples available many cases, which leads failure conventional methods. Clustering incomplete referred clustering. In view promising application prospects, research has noticeable advances recent years. there no...
In this paper, we propose a novel method, referred to as incomplete multi-view tensor spectral clustering with missing-view inferring (IMVTSC-MVI) address the challenging problem missing views. Different from existing methods which commonly focus on exploring certain information of available views while ignoring both hidden and intra-view data, IMVTSC-MVI seeks recover explore full such recovered for data clustering. particular, incorporates feature space based manifold similarity graph...
Bounding box regression (BBR) has been widely used in object detection and instance segmentation, which is an important step localization. However, most of the existing loss functions for bounding cannot be optimized when predicted same aspect ratio as groundtruth box, but width height values are exactly different. In order to tackle issues mentioned above, we fully explore geometric features horizontal rectangle propose a novel similarity comparison metric MPDIoU based on minimum point...
Video anomaly detection (VAD) refers to the discrimination of unexpected events in videos. The deep generative model (DGM)-based method learns regular patterns on normal videos and expects learned yield larger errors for abnormal frames. However, DGM cannot always do so, since it usually captures shared between events, which results similar them. In this article, we propose a novel self-supervised framework unsupervised VAD tackle above-mentioned problem. To end, design attentive adversarial...