Adaptive plane orthogonal to skin-tone vector for non-contact pulse rate monitoring in motion scenarios

This study proposes a novel method to replace the fixed plane orthogonal to the skin-tone vector in the Plane-Orthogonal-to-Skin (POS) method to obtain a more accurately pulse rate in motion scenarios. To evaluate the performance of the proposed method, we constructed a series of 155 challenge videos from 26 subjects, which included static scenarios, motion scenarios, dark scenarios, bright scenarios, long-distance scenarios, and after exercise scenarios. The experimental results show that this method can achieve high-precision pulse rate results with mean absolute error can be less than 5.51 bpm. Compared to current mainstream motion suppression methods (e.g. POS, Chrominance (CHROM), Independent Component Analysis (ICA), Local Group Invariance (LGI)), our method shows the best performance. As a result, it has enormous potential for intelligent health monitoring.