Chong Tang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Chong Tang has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Guang Feng
School of Computer Science and Technology, Guangdong University of Technology (GDUT), Guangzhou, China
2 shared publications - Qun Yang
College of Automation, Guangdong University of Technology, Guangzhou 511400, China
1 shared publication - Wenyan Wu
School of Computer Science, Guangdong University of Technology, Guangzhou, China
1 shared publication - Xiaoting Wu
College of Automation, Guangdong University of Technology, Guangzhou 511400, China
1 shared publication - Yunhai Liu
College of Computer, Guangdong University of Technology, Guangzhou 511400, China
1 shared publication
Latest Publications
- Open Access
- 24 April 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (9) , 3606
Mask-Wearing Detection in Complex Environments Based on Improved YOLOv7
Wearing masks is an effective protective measure for residents to prevent respiratory infectious diseases when going out. Due to issues such as a small target size, target occlusion leading to information loss, false positives, and missed detections, the effectiveness of face mask-wearing detection needs improvement. To address these issues, an improved YOLOv7 object detection model is proposed. Firstly, the C2f_SCConv module is introduced in the backbone network to replace some ELAN modules for feature extraction, enhancing the detection performance of small targets. Next, the SPPFCSPCA module is proposed to optimize the spatial pyramid pooling structure, accelerating the model convergence speed and improving detection accuracy. Finally, the HAM_Detect decoupled detection head structure is introduced to mitigate missed and false detections caused by target occlusion, further accelerating model convergence and improving detection performance in complex environments. The experimental results show that improved YOLOv7 achieved an mAP of 90.1% on the test set, a 1.4% improvement over the original YOLOv7 model. The detection accuracy of each category improved, effectively providing technical support for mask-wearing detection in complex environments.
- 24 April 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (9) , 3606
Open AccessMask-Wearing Detection in Complex Environments Based on Improved YOLOv7
Wearing masks is an effective protective measure for residents to prevent respiratory infectious diseases when going out. Due to issues such as a small target size, target occlusion leading to information loss, false positives, and missed detections, the effectiveness of face mask-wearing detection needs improvement. To address these issues, an improved YOLOv7 object detection model is proposed. Firstly, the C2f_SCConv module is introduced in the backbone network to replace some ELAN modules for feature extraction, enhancing the detection performance of small targets. Next, the SPPFCSPCA module is proposed to optimize the spatial pyramid pooling structure, accelerating the model convergence speed and improving detection accuracy. Finally, the HAM_Detect decoupled detection head structure is introduced to mitigate missed and false detections caused by target occlusion, further accelerating model convergence and improving detection performance in complex environments. The experimental results show that improved YOLOv7 achieved an mAP of 90.1% on the test set, a 1.4% improvement over the original YOLOv7 model. The detection accuracy of each category improved, effectively providing technical support for mask-wearing detection in complex environments.
- Open Access
- 9 February 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (4) , 1439
Portrait Semantic Segmentation Method Based on Dual Modal Information Complementarity
Semantic segmentation of human images is a research hotspot in the field of computer vision. At present, the semantic segmentation models based on U-net generally lack the ability to capture the spatial information of images. At the same time, semantic incompatibility exists because the feature maps of encoder and decoder are directly connected in the skip connection stage. In addition, in low light scenes such as at night, it is easy for false segmentation and segmentation accuracy to appear. To solve the above problems, a portrait semantic segmentation method based on dual-modal information complementarity is proposed. The encoder adopts a double branch structure, and uses a SK-ASSP module that can adaptively adjust the convolution weights of different receptor fields to extract features in RGB and gray image modes respectively, and carries out cross-modal information complementarity and feature fusion. A hybrid attention mechanism is used in the jump connection phase to capture both the channel and coordinate context information of the image. Experiments on human matting dataset show that the PA and MIoU coefficients of this algorithm model reach 96.58% and 94.48% respectively, which is better than U-net benchmark model and other mainstream semantic segmentation models.
- 9 February 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (4) , 1439
Open AccessPortrait Semantic Segmentation Method Based on Dual Modal Information Complementarity
Semantic segmentation of human images is a research hotspot in the field of computer vision. At present, the semantic segmentation models based on U-net generally lack the ability to capture the spatial information of images. At the same time, semantic incompatibility exists because the feature maps of encoder and decoder are directly connected in the skip connection stage. In addition, in low light scenes such as at night, it is easy for false segmentation and segmentation accuracy to appear. To solve the above problems, a portrait semantic segmentation method based on dual-modal information complementarity is proposed. The encoder adopts a double branch structure, and uses a SK-ASSP module that can adaptively adjust the convolution weights of different receptor fields to extract features in RGB and gray image modes respectively, and carries out cross-modal information complementarity and feature fusion. A hybrid attention mechanism is used in the jump connection phase to capture both the channel and coordinate context information of the image. Experiments on human matting dataset show that the PA and MIoU coefficients of this algorithm model reach 96.58% and 94.48% respectively, which is better than U-net benchmark model and other mainstream semantic segmentation models.