Finger Vein Recognition Based on ResNet With Self-Attention
To solve the problem of low accuracy and high computational resource consumption in finger vein recognition, a finger vein recognition model based on ResNet with self-attention (FV-RSA) is proposed. This model combines global focusing ability of self-attention mechanism and local feature extraction...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10375376/ |
| _version_ | 1797361438409359360 |
|---|---|
| author | Zhibo Zhang Guanghua Chen Weifeng Zhang Huiyang Wang |
| author_facet | Zhibo Zhang Guanghua Chen Weifeng Zhang Huiyang Wang |
| author_sort | Zhibo Zhang |
| collection | DOAJ |
| description | To solve the problem of low accuracy and high computational resource consumption in finger vein recognition, a finger vein recognition model based on ResNet with self-attention (FV-RSA) is proposed. This model combines global focusing ability of self-attention mechanism and local feature extraction ability of CNN, which improves recognition accuracy. To reduce the number of parameters and floating-point operations, self-attention and convolution share linear projections by pointwise convolution. Self-attention and CNN are fused in the convolution and self-attention (CASA) block connected by skip connection to avoid gradient vanishing or gradient exploding. During the training phase, we use a variable learning rate with cosine annealing to avoid falling into local optimum. Experiments show that the method works well on the public database, which can not only improve the accuracy, but also reduce the number of parameters and computational complexity. |
| first_indexed | 2024-03-08T15:54:45Z |
| format | Article |
| id | doaj.art-7b847b593dec4000804eb3b8546b59ac |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-03-08T15:54:45Z |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-7b847b593dec4000804eb3b8546b59ac2024-01-09T00:04:46ZengIEEEIEEE Access2169-35362024-01-01121943195110.1109/ACCESS.2023.334792210375376Finger Vein Recognition Based on ResNet With Self-AttentionZhibo Zhang0https://orcid.org/0000-0001-8830-3935Guanghua Chen1https://orcid.org/0000-0002-7963-0974Weifeng Zhang2https://orcid.org/0009-0003-8337-1826Huiyang Wang3https://orcid.org/0009-0008-1525-1263School of Mechatronic Engineering and Automation, Shanghai University, Shanghai, ChinaSchool of Mechatronic Engineering and Automation, Shanghai University, Shanghai, ChinaCollege of Electromechanical Engineering, Qingdao University of Science and Technology, Qingdao, ChinaSchool of Mechatronic Engineering and Automation, Shanghai University, Shanghai, ChinaTo solve the problem of low accuracy and high computational resource consumption in finger vein recognition, a finger vein recognition model based on ResNet with self-attention (FV-RSA) is proposed. This model combines global focusing ability of self-attention mechanism and local feature extraction ability of CNN, which improves recognition accuracy. To reduce the number of parameters and floating-point operations, self-attention and convolution share linear projections by pointwise convolution. Self-attention and CNN are fused in the convolution and self-attention (CASA) block connected by skip connection to avoid gradient vanishing or gradient exploding. During the training phase, we use a variable learning rate with cosine annealing to avoid falling into local optimum. Experiments show that the method works well on the public database, which can not only improve the accuracy, but also reduce the number of parameters and computational complexity.https://ieeexplore.ieee.org/document/10375376/Finger vein recognitiondeep learningResNetself-attentionvariable learning rate |
| spellingShingle | Zhibo Zhang Guanghua Chen Weifeng Zhang Huiyang Wang Finger Vein Recognition Based on ResNet With Self-Attention IEEE Access Finger vein recognition deep learning ResNet self-attention variable learning rate |
| title | Finger Vein Recognition Based on ResNet With Self-Attention |
| title_full | Finger Vein Recognition Based on ResNet With Self-Attention |
| title_fullStr | Finger Vein Recognition Based on ResNet With Self-Attention |
| title_full_unstemmed | Finger Vein Recognition Based on ResNet With Self-Attention |
| title_short | Finger Vein Recognition Based on ResNet With Self-Attention |
| title_sort | finger vein recognition based on resnet with self attention |
| topic | Finger vein recognition deep learning ResNet self-attention variable learning rate |
| url | https://ieeexplore.ieee.org/document/10375376/ |
| work_keys_str_mv | AT zhibozhang fingerveinrecognitionbasedonresnetwithselfattention AT guanghuachen fingerveinrecognitionbasedonresnetwithselfattention AT weifengzhang fingerveinrecognitionbasedonresnetwithselfattention AT huiyangwang fingerveinrecognitionbasedonresnetwithselfattention |