Optical Instruments, Volume. 46, Issue 6, 55(2024)
Fine-grained image classification algorithm combining saliency and non-local module
Figures & Tables(8)
Table 1. Classification accuracy on Stanford Cars and CUB-200-2011
View tableView in ArticleTable 1. Classification accuracy on Stanford Cars and CUB-200-2011
算法 准确率/% Stanford Cars CUB-200-2011 注：黑体表示最优结果 Cutout 92.12 84.61 Mixup 92.49 84.80 RICAP 92.27 84.66 Gridmask 92.07 84.27 FMix 92.15 84.71 SaliencyMix 92.30 85.04 B-CNN 90.60 84.00 RA-CNN 92.50 85.30 MC-Loss 91.30 84.30 CNL 93.10 85.80 本方法 94.01 85.97
Table 2. Experimental result of ablation on CUB-200-2011
View tableView in ArticleTable 2. Experimental result of ablation on CUB-200-2011
模块及方法 准确率/% Stage2 Stage3 注：黑体表示最优结果 ResNet-50 83.85 83.85 +1非局部 84.70 84.83 +2非局部 84.91 85.07 +3非局部 85.05 85.21 +增强 85.19 85.19 +增强+1非局部 85.30 85.37 +增强+2非局部 85.41 85.50 +增强+3非局部 85.46 85.97
Tools
Get Citation
Copy Citation Text
Chen LING, Rongfu ZHANG, Ziye YANG, Guyu GAO, Fuqiang ZHAO. Fine-grained image classification algorithm combining saliency and non-local module[J]. Optical Instruments, 2024, 46(6): 55
Paper Information
Category:
Received: Dec. 5, 2023
Accepted: --
Published Online: Jan. 21, 2025
The Author Email: ZHANG Rongfu (zrf@usst.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20