Video Magnification and Photoplethysmography Based Heart Rate Measurement System

Magnification of subtle temporal changes like motion or color is known as video magnification. This paper explores Eulerian method for video magnification. It involves Gaussian and Box filters magnifying motion and color variations in video respectively. This algorithm has been applied to detect human heart rate by obtaining Blood Volume Pulse (BVP) signal from magnified video of finger. The finger video is obtained using low-cost mobile camera. The heart rate obtained by our method is very close to clinical heart rate obtained using Phillips multipara monitor. Preprocessing involves the converting the video into frames array. After this conversion of the video into frames, an individual frame goes through the noise removal process. Video-frames may contain many types of noise that directly affect the result of the video magnification. Motion denoising is one of the techniques that deal with motion noise in the video frames. Further processing continues with the motion and color magnification which is a main part of a video magnification algorithm. In this work, Eulerian and Lagrangian motion magnification methods are implemented. This helps in ameliorating the final results of the video magnification process. Algorithm revel subtle changes in the video which are visible to the necked eye and is useful in analyzing vital signs of person, like observing the facial blood circulation. It is also useful in monitoring object subtle movement under various constraints.