Optics and Precision Engineering, Volume. 31, Issue 21, 3192(2023)
Face recognition algorithm incorporating CBAM and Siamese neural network
Figures & Tables(13)
Fig. 5. Model structure diagram of VGG11_BN combined with Siamese neural network
Fig. 9. Comparison between proposed algorithm and traditional Siamese neural network
Table 1. Comparison of different learning rates in proposed algorithm
View tableView in ArticleTable 1. Comparison of different learning rates in proposed algorithm
LR AUC 平均准确率（Acc） 平均精度（mAP） Triangular 98.22 92.34 98.11 Triangular2 99.36 96.67 99.37 Exp_range 95.76 89.54 93.58
Table 2. Comparison of accuracy rates of different algorithms
View tableView in ArticleTable 2. Comparison of accuracy rates of different algorithms
Method Recognition rate LWSFLA-SVM［ 26 ］93.00 SRGES［ 27 ］88.25 KEHD ［ 28 ］92.17 KCCA-SVM［ 29 ］93.50 VGG16+siamese 85.50 VGG11+siamese 87.60 Proposed 94.30
Table 3. Results of ablation experiment
View tableView in ArticleTable 3. Results of ablation experiment
模 型 计算量
/GFLOPs
平均准确率/% CASIA-FaceV5 CAS-PEAL-R1 Base 15.06 81.40 87.6 CBAM+Base 12.60 85.90 87.5 Base+CBAM 14.80 90.30 92.2 本文算法 13.50 96.67 94.3
Tools
Get Citation
Copy Citation Text
Xiangzhou MENG, Yingjun LI, Guicong WANG, Tiansheng MENG. Face recognition algorithm incorporating CBAM and Siamese neural network[J]. Optics and Precision Engineering, 2023, 31(21): 3192
Paper Information
Category:
Received: May. 30, 2023
Accepted: --
Published Online: Jan. 5, 2024
The Author Email: LI Yingjun (me_liyj@ujn.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20