GPU-Accelerated Non-Contact Heart Rate Monitoring for Biomedical Applications*

Non-contact heart rate monitoring has emerged as a promising approach for biomedical applications, offering advantages in hygiene, convenience, and scalability. However, real-time processing of physiological signals from multiple subjects remains computationally challenging. This paper presents a GPU-accelerated real-time non-contact heart rate monitoring system designed for multi-subject scenarios. By leveraging parallel computing on GPU, our method efficiently processes remote photoplethysmography (rPPG) signals extracted from facial or skin regions through video analysis and apply signal conditioning and spectral analysis to estimate the heart rate and its trend. Experimental results demonstrate that the proposed system achieves real-time multi-person monitoring, simultaneously tracking over 10 subjects with GPU acceleration delivering a 10x speedup compared to CPU implementations. Notably, the system maintains robust correlation coefficients (R > 0.86) even in challenging scenarios, demonstrating reliable heart rate trend extraction.