A multimodal liveness detection method based on palmprint features and remote photoplethysmography signals and its performance evaluation

Objective To investigate a novel liveness detection method by fusing palmprint features with photoplethysmography (PPG) signals, so as to enhance the anti-spoofing capability of palm-based biometric systems and evaluate its detection accuracy and applicability. Methods A multimodal liveness detection scheme integrating palmprint and PPG characteristics was designed. A high-definition camera was used to capture experimental samples under various lighting conditions and spoofing attacks. An approach combining image processing and temporal signal analysis was employed to extract PPG signals from palm videos while concurrently extracting palmprint features for identity verification. The performance of the method was systematically evaluated regarding signal extraction accuracy, stability, anti-interference capability, and liveness detection accuracy. Results Experimental validation across diverse samples and spoofing attack conditions demonstrated that the method could stably extract PPG signals. The signal strength showed a high correlation with signals obtained from standard measurement devices. The results indicated that the proposed multimodal liveness detection method, based on the fusion of palmprint features and PPG signals, could effectively identify liveness characteristics under different physiological states, thus achieving a liveness detection accuracy rate of over 90%. Conclusion This study proposes a multimodal liveness detection method based on the fusion of palmprint features and PPG signals. By integrating static palmprint information with dynamic PPG signals, it significantly enhances the security, stability, and robustness of palm-based biometric recognition systems.