A5100341679- Industrial Vision Systems and Defect Detection
- Image Processing Techniques and Applications
- Visual Attention and Saliency Detection
- Advanced Neural Network Applications
- Advanced Steganography and Watermarking Techniques
- Remote-Sensing Image Classification
- Image and Object Detection Techniques
- Digital Media Forensic Detection
- Biometric Identification and Security
- Manufacturing Process and Optimization
- Advanced Image and Video Retrieval Techniques
- Advanced Measurement and Detection Methods
- Smart Agriculture and AI
- Infrared Target Detection Methodologies
- Optical measurement and interference techniques
- Advanced Image Fusion Techniques
- Surface Roughness and Optical Measurements
- Chaos-based Image/Signal Encryption
- User Authentication and Security Systems
- Textile materials and evaluations
- Emotion and Mood Recognition
- Face recognition and analysis
- Face and Expression Recognition
- Product Development and Customization
- Image and Signal Denoising Methods
Zhongyuan University of Technology
2016-2025
Zhengzhou University of Light Industry
2024
Hohai University
2024
China University of Petroleum, East China
2023
Research Institute of Petroleum Exploration and Development
2023
Xiamen University
2023
Baoji University of Arts and Sciences
2021-2023
Shangdong Agriculture and Engineering University
2021
China Academy of Engineering Physics
2020-2021
Anhui University
2021
To improve the accuracy of color image completion with missing entries, we present a recovery method based on generalized higher-order scalars. We extend traditional second-order matrix model to more comprehensive equivalent, called “t-matrix” model, which incorporates pixel neighborhood expansion strategy characterize local constraints. This is then used some commonly and tensor algorithms their versions. perform extensive experiments various using simulated data publicly available images....
Facial expression recognition (FER) in the wild is challenging due to various unconstrained conditions, i.e., occlusions and head pose variations. Previous methods tend improve performance of facial through resorting holistic or coarse local-based methods, while ignoring local fine-grained feature structure knowledge correlation between features. In this paper, we propose a Fine-Grained Association Graph Representation (FG-AGR) framework which can capture representation. Firstly, an Adaptive...
In contrast to defect-free fabric images with macro-homogeneous textures and regular patterns, the defect are characterized by regions that salient sparse among redundant background. Therefore, as an effective tool for separating image into a part (the background) defect), low-rank decomposition model provides ideal solution patterned detection. this paper, novel method detection is proposed based on texture descriptor model. First, efficient second-order orientation-aware descriptor,...
In this paper, a novel fabric detect detection scheme based on HOG and SVM is proposed. Firstly, each block-based feature of the image encoded using histograms orientated gradients (HOG), which are insensitive to various lightings noises. Then, powerful selection algorithm, AdaBoost, performed automatically select small set discriminative features in order achieve robust results. end, support vector machine (SVM) used classify defects. Experimental results demonstrate efficiency our proposed...
Fabric defect detection plays a key role in the quality control of textiles. Existing fabric methods adopt traditional pattern recognition methods; however, these lack adaptability and present poor performance. Because biological vision system has ability to quickly locate salient objects, we propose novel algorithm based on modeling by simulating mechanism visual perception. First, distinct, efficient, robust feature descriptor from P ganglion cells, which was proposed our previous work, is...
Solid and accurate object detection in optical remote sensing images still remains significant challenges such as complex background weak information. To alleviate above problems, we propose a revolutionary one-stage network. Specifically, the proposed effective localization attention is embedded deep feature maps with more channels, used to locate channels that are for tasks through one dimensional convolution operations. Following that, small compensation strategy use fusion operation...
Advancements in deep neural networks have made remarkable leap-forwards crop detection. However, the detection of wheat ears is an important yet challenging task due to complex background, dense targets, and overlaps between ears. Currently, many detectors significant progress improving accuracy. some them are not able make a good balance computational cost precision meet needs deployment real world. To address these issues, lightweight efficient ear detector with Shuffle Polarized...
Fabric defect recognition is an important measure for quality control in a textile factory. This article utilizes deep convolutional neural network to recognize defects fabrics that have complicated textures. Although networks are very powerful, large number of parameters consume considerable computation time and memory bandwidth. In real-world applications, however, the fabric task needs be carried out timely fashion on computation-limited platform. To optimize network, novel method...
In this paper, Fabric defect detection is a challenging task because of the complex texture. Deep learning technology provide promising solution. As kind deep object model. Single Shot Multibox Detector(SSD)achieves good performance. However, original SSD model may fail to detect small objects. we proposed novel for fabric detection. Experimental results showed that improved can accurately region.
Surface defect detection (SDDet) is extremely critical for quality control and routine maintenance, computer vision-based methods have delivered promising performance in various industrial fields. However, there remains big challenges region consistency boundary localization, due to complicated appearance defects, interference of artifacts, low contrast. To overcome these issues, this article, we propose an enhanced encoder–decoder network with hierarchical supervision SDDet. Specifically,...
Fabric defect detection is crucial in the textile industry, as it suffers from challenges such small sizes, diverse morphologies, and imbalanced sample distributions. Current mainstream methods approach an object problem. Many fabric defects, particularly ones, are caused by production faults that disrupt texture. These defects often lack distinct structural information, which poses a challenge for rely heavily on features. To address these limitations, we propose multi-scale channel based...
Fabric defect detection is one of the hot topics in textile industry, primarily due to occasional machine failures and repetitive errors on production lines leading misalignment warp weft. Consequently, collecting samples challenging, as these defects appear subtle, linear shapes horizontal or vertical orientations, resulting sparse features with low contrast. This poses significant challenges model’s feature learning capabilities performance. we developed a research platform for developing...