SU-E-I-73: Visible Image Sequence Based Bio Signal Extraction System and Algorithm for Personal Healthcare

Abstract

Purpose: In this paper, we introduced the new method to extract the heart rate using Fourier Descriptor and video sequence images. Remote measurements of the heart rate can provide comfortable physiological assessment without external sensors. This novel approach can be applied to color video recordings of the human face and is based on automatic face segmentation along with Otsu method. This is the demonstration of a lowcost accurate video-based method for contact-free heart rate measurements that is automated. Methods: In this study, we propose an algorithm for analyzing the sequence of video obtained from the mobile phone, to measure the number of beating of the heart. After For this experiment, was seated on a chair in the first experimenter, and simultaneously imaging the portion of the face in the cell phone camera, was measured PPG. Classifying each frame, by using the face recognition system, a video sequence is acquired, was segmentation portion of the face of optimal. Then, divide each frame red channel of the image, green channel, the blue channel, it was added the values of all pixels in the green channel, and obtain an average by dividing the value of each pixel. After obtained using the Fourier descriptor frequencies were transformed using a fast Fourier transform with maximum amplitude, the change values of the pixels of the Green channel is converted to a per minute heart rate frequency. Results: From the experiments, the error between PPG value and the heart rate estimation value from the image sequence are 1.39% of average relative error, 8.12% of maximum relative error and 0.14% of minimum relative error. Conclusion: In this paper, we studied the pulse measurement algorithm of sequence-based video using the Fourier Descriptor in this paper.