Research Of Heart Rate Detection Based On Facial Video

Heart rate is the number of heartbeats per unit of time, usually per minute. It is one of the most important and basic physiological parameters that used to evaluate physical condition of a human. At present, there are a lot of methods for detecting heart rate. All those methods can be divided into contact detection method and non-contact detection method according to the process of heart rate detection. Contact heart rate detection method such as Electrocardiography(ECG), finger clip or chest strap needs a direct contact with patients body to measure heart rate. Non-contact measurements can provide comfortable physiological assessment without clutter of wiring and electrodes. A new method of detecting heart rate, which based on photoplethysmographic imaging, is studied in this paper. Computer vision technology and Eulerian Video Magnification are used in our procedure to extract heart rate signal from a video of human face. The main tasks and results of this paper can be summarized as follows:1. Capture vide of human face. A video of a human face is captured under ambient sunlight. Then face detection method is used to locate the position of human face in each frame of the video. Finally a region of the face location is selected as ROI which will be used to extract heart rate signal.2. Eulerian Video Magnification. Eulerian Video Magnification first decomposes the input video sequence into different spatial frequency bands, and applies the same temporal filter to all bands. Apply Eulerian Video Magnification to vide of human face to magnify weak heart rate signal.3. Consistency between the proposed method in this paper and the tradition method. To illustrate the accuracy of our method, we use Bland-Altman method to access the consistency between the proposed method and the traditional method.4. Analysis of factors that affect the accuracy of heart rate detection. We analysis the primary factors that affect the accuracy of heart rate detection according to the characteristic of the method proposed in this paper. The effect of different color model and different ROI are analyzed.The method proposed in this paper is used to detect the heart rate of five participants through45videos of face captured at different times of a day. The result shows that the proposed method is accurate and effective.