Memory Assisting System for Elderly

Abstract

In Hong Kong, there is about 10% elderly suffering from alzheimer’s disease. To help the patients, the memory assisting system tries to reinforce elderly memories and prevent them from heavy stress. User of the system can take photo of a person anytime and show his or her identity immediately. The project intends to develop an accurate face recognition system running on mobile app. According to the tests of different pre-processing and recognition methods, the best combination had been taken to build up the unique, precise enough system. The design of the system was divided into three parts including pre-processing part, face detection part and face recognition part. Pre-processing part aims to provide higher quality of input photos and ensure consistency between training photos as well as query photo. Face detection part can detect all the facial features. Face recognition part includes the database and particular matching method. Due to time limitation, the system now is running on Window platform. User can input photo into the system. It can recognize and show the identity of the person in the photo. It was found that the system provide high accuracy of face detection and recognition. The result is acceptable. Future work should be done to achieve our final target and contribute a more all-rounded, precise system.