A5100398796- Industrial Vision Systems and Defect Detection
- Surface Roughness and Optical Measurements
- Image Processing Techniques and Applications
- Advanced Neural Network Applications
- Manufacturing Process and Optimization
- Photovoltaic System Optimization Techniques
- Image and Object Detection Techniques
- Non-Destructive Testing Techniques
- Optical measurement and interference techniques
- Structural Load-Bearing Analysis
- Structural Behavior of Reinforced Concrete
- Remote Sensing and Land Use
- Integrated Circuits and Semiconductor Failure Analysis
- Advanced machining processes and optimization
- Advanced Manufacturing and Logistics Optimization
- Radiomics and Machine Learning in Medical Imaging
- Infrastructure Maintenance and Monitoring
- Advanced Algorithms and Applications
- Advanced Measurement and Metrology Techniques
- Advanced Fiber Optic Sensors
- Industrial Technology and Control Systems
- Photovoltaic Systems and Sustainability
- Advanced Measurement and Detection Methods
- EEG and Brain-Computer Interfaces
- Advanced Fiber Laser Technologies
Hebei University of Technology
2016-2025
China Mobile (China)
2024
Qingdao University of Science and Technology
2024
University of Buenos Aires
2024
Natural Resources Canada
2024
Beijing University of Technology
2015-2023
Hunan University of Technology
2023
Civil Aviation University of China
2023
Technische Universität Ilmenau
2023
Second People Hospital of Hunan
2023
The automatic defects detection for solar cell electroluminescence (EL) images is a challenging task, due to the similarity of defect features and complex background features. To address this problem, in article novel complementary attention network (CAN) designed by connecting channel-wise subnetwork with spatial sequentially, which adaptively suppresses noise highlights simultaneously employing advantage channel position In CAN, applies convolution operation integrate concatenated...
Automatic defect detection on the steel surface is a challenging task in computer vision, owing to miscellaneous patterns of defects, low contrast between and background, existence pseudo so on. In this paper, new Haar-Weibull-variance (HWV) model proposed for an unsupervised manner. First, anisotropic diffusion utilized eliminate influence pseudodefects. Second, HWV established characterize texture distribution each local patch image. The can project into low-dimensional space with only two...
Automatic defect detection on strip steel surfaces is a challenging task in computer vision, owing to miscellaneous patterns of defects, disturbance pseudodefects, and random arrangement gray-level background. In this paper, novel template establishment presented. Further, simple guidance template-based algorithm for surface proposed. First, large number defect-free images are collected obtain the statistical characteristic normal textures. Second, each given test image, initial built...
The small hot spot defect detection for photovoltaic (PV) farms is a challenging problem due to the feature vanishing as network deepens. To solve this problem, novel residual channelwise attention gate (RCAG-Net) proposed by employing RCAG module achieve multiscale fusion, complex background suppression, and highlighting. In RCAG-Net, first realizes fusion adding features of different scale layers. Next, global average pooling (GAP) multilayer perceptron (MLP) are used dimension reduction...
Automatic strip steel surface defect detection is a difficult mission, as result of the imbalanced class distributions caused by sparse distribution abnormal samples. The one-class classification (OCC) method can detect samples only training normal Generative Adversarial Networks (GAN) automatically learn features in unsupervised situations, and one sample used to train model. GAN-based for defects proposed paper. second last output layer GAN generator chosen feature, which contains some...
Automatic vision-based defect detection on the steel surface is a challenging task due to miscellaneous patterns of defects, low contrast between and background, so on. Image-decomposition-based method can analyze structure texture inspect defective objects. Currently, state art image decomposition-based methods one guided by given fixed template. However, template cannot be suitable for all situations. In this article, new self-reference template-guided decomposition algorithm strip...
Neural Representations for Videos (NeRV) has emerged as a promising implicit neural representation (INR) approach video analysis, which represents videos networks with frame indexes inputs. However, NeRV-based methods are time-consuming when adapting to large number of diverse videos, each requires separate NeRV model be trained from scratch. In addition, spatially require generating high-dimension signal (i.e., an entire image) the input low-dimension timestamp, and typically consists tens...
Vibration signal analysis is one of the most effective methods for mechanical fault diagnosis. Available part information always concealed in component noise, which makes it much more difficult to detect defection, especially at early stage development. This paper presents a new approach diagnosis based on time domain and adaptive fuzzy<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M2"><mml:mrow><mml:mi>C</mml:mi></mml:mrow></mml:math>-means clustering. By analyzing vibration...
Abstract Major depressive disorder (MDD) is one of the most common psychiatric disorders worldwide with high recurrence rate. Identifying MDD patients, particularly those recurrent episodes resting‐state fMRI, may reveal relationship between and brain function. We proposed a Transformer‐Encoder model, which utilized functional connectivity extracted from large‐scale multisite rs‐fMRI datasets to classify HC. The model discarded Transformer's Decoder part, reducing model's complexity...
The surface of a multicrystal solar cell shows multiple crystal grains random shapes and sizes. It creates an inhomogeneous texture in the surface, which brings great difficulty to automatic crack detection polycrystalline surface. As perceptual grouping approach, tensor voting can extract curvilinear structures such as lines curves from noisy, binary data 2-D or 3-D, without invoking specific object model. However, traditional be susceptible gap problem structural noise. To address problems...
Due to the multi-scale characteristics of defects and strong background interference, automation solar cell surface defect detection is still a challenge. To address this problem, paper proposes novel ohject detector called BPGA-Detector which consists two parts: Bidirectional-Path Feature Pyramid Network (BPFPN) Group-wise Attention Module (GAM). The combines multiscale features using bidirectional-path feature fusion method that structured by connecting bottom-up path pyramid network...