Mobile Device and Cloud Server based Intelligent Health Monitoring Systems

Abstract

Health awareness of people is generally improved in the past decade. On one hand, with the accelerating pace of modern life, non-intrusive intelligent health monitoring becomes desirable. On the other hand, the Android smart phones equipped with various sensors and powerful processing unit are increasingly popular in the market. In this project, I programmed an Android App capable of issuing health alerts based on audio and visual processing. Raw data was obtained from the built-in camera and the microphone. Signal processing and machine learning algorithms were applied to give intelligent feedback. The first part of the App is to calculate the user's speech pitch at run time and to check for speech disorders. The second function is to measure the user's heart rate using real time fingertip image processing. The last feature is to classify the user's emotion status from the captured facial image and to record the result on database for mental condition monitoring. Health monitoring on mobile devices is intrinsically challenging due to its complex nature and hardware limitation. Satisfactory recognition accuracy has been achieved in the development stage. Higher accuracy is expected after the App is released as beta version and tested with larger training dataset.