Hand Multimodal Recognition Based on Adaptive Multichannel Graph Convolutional Networks
Xinbo Lai, Alimjan Abdireyim, Reyihanguli Kasenmu, Nurbiya Yadikar, Kurban Ubul · 2024
In the biometrics field, the unimodal biometrics recognition effect cannot meet the requirements for high-performance identity recognition due to its characteristics, such as instability and limitations. Multimodal biometrics combines two or more biometric features to improve the accuracy and security of identification, and combining multiple biometric features can make up for the shortcomings of unimodal biometrics. Aiming at the above problems, this paper proposes a hand multimodal recognition network based on an adaptive multichannel graph convolutional network, which adopts the PageRank centrality fusion method for image fusion of fingerprints, finger veins, and palm veins so that the network takes into account both the feature novelty information and the structural information in the image fusion process. At the same time, we introduce the CBAM attention module in the network to allow the network to adaptively learn important features in channel and spatial dimensions to improve the network's performance. The experimental results show that the recognition accuracy of the method proposed in this paper reaches 99.96%, which has high performance and can be widely used in practical applications.