ResNeXt with Cutout for Finger Vein Analysis
Jeffrey Yeh, Hung-Tse Chan, Chih‐Hsien Hsia · 2021 International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS) · 2021
In recent years, Convenience and security are of increasing importance so traditional keys and password locks are being replaced because they can be forgotten or lost, deciphered or stolen. Various types of personal identification systems that use biometrics are secure, reliable and convenient. This study uses the cutout for data augmentation and the ResNeXt as a backbone. The experimental results show that the proposed method has an accuracy of 98.10% for the FV-USM public database.