CNN-Based Human Identification Method Using Image Photoplethysmographic

Biometrics has received extensive attention due to its accuracy and convenience. However, commonly used biometric features still have deficiencies. Facial recognition is a potential threat to privacy, and iris recognition or heartbeat recognition requires specific acquisition equipment, resulting in additional costs. To address this issue, we proposed a novel biometric identification method using image photoplethysmographic (IPPG). IPPG signal is easy collection with a consumer camera and the pixel-averaging operation to extract IPPG signal will remove private facial information. As a vital sign, IPPG signal is difficult to fake using abiotic prostheses. We constructed a CNN-based identification with IPPG (ID-IPPG) to verify the performance of IPPG signals in human identification. The proposed model achieves 97.3% accuracy on IPPG signals dataset containing 12 subjects. Moreover, the model can effectively perform in living body detection. The results demonstrate that IPPG signals contain individual physiological information and the ID-IPPG has high accuracy and security for human identification.