Estimating heart rate via depth video motion tracking

摘要: Depth sensors like Microsoft Kinect can acquire partial geometric information in a 3D scene via captured depth images, with potential application to non-contact health monitoring. However, videos typically suffer from low bit-depth representation and acquisition noise corruption, hence using them deduce metrics that require tracking subtle structural details is difficult. In this paper, we propose capture video 2.0 estimate the heart rate of human subject; as blood pumped circulate through head, tiny oscillatory head motion be detected for periodicity analysis. Specifically, first perform joint enhancement / denoising procedure improve quality graph-signal smoothness prior regularization. We then track an automatically nose region throughout vectors. The deduced vectors are analyzed principal component analysis rate. Experimental results show improved accuracy our proposed procedure, estimated rates close ground truth.