Heart Rate Estimation based on Remote Imaging PPG for Android Mobile Devices

Abstract

In this project, we need to design and implement an Android application which can measure the human heart rate by scanning the face and analysing the video. It is based on RIPPG (Remote Imaging Photoplethysmography) Theory, which is used colour changes in cheeks skin during the pumping action of heart to measure the heart rate. To build the application, firstly, we need to import OpenCV Library and combine C++ language with Java by using Java Native Interface (JNI) and Android NDK in this application. Then, face detection is used to locate the cheeks of users and calculate the colour differences in Red and Green channels in those regions. Lastly, after applying Fast Fourier Transformation on the data of colour differences, we get the frequency with the largest power spectrum, which is the value of user’s heart rate at that moment. The result of this application is reasonable and it could give the heart rate in the correct interval. Furthermore, since the large computation in face detection which caused some delays in video processing, I tried to apply face tracking algorithm instead of face detection to obtain a more smooth video and reduce some noises in data in order to improve the accuracy of the result.