Laser & Optoelectronics Progress, Volume. 61, Issue 10, 1037004(2024)
Multi-spectral Pedestrian Detection Based on Deformable Convolution and Multi-Scale Residual Attention
Fig. 1. Sampling position. (a) Regular convolution; (b) DCN
Fig. 2. 3×3 DCN
Fig. 3. Schematic diagram of SPP module
Fig. 4. Schematic diagram of multi-scale residual attention module. (a) Overall structure; (b) structure of UP mudule
Fig. 5. Neck before improvement
Fig. 6. Neck after improvement
Fig. 7. Improved YOLOv5s network
Fig. 8. Curves of P-R
Fig. 9. Visualization comparison of detection results by different algorithms. (a) Complex background scene; (b) heavily occlusion pedestrian target scene; (c) blur background small target scene; (d) weak feature pedestrian target scene
|
|
|
|
Get Citation
Copy Citation Text
Guoli Zhang, Shuai Chang, Yansong Song, Tianci Liu. Multi-spectral Pedestrian Detection Based on Deformable Convolution and Multi-Scale Residual Attention[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1037004
Category: Digital Image Processing
Received: Sep. 15, 2023
Accepted: Oct. 20, 2023
Published Online: Mar. 20, 2024
The Author Email: Chang Shuai (changshuai@cust.edu.cn)