ASSISTED NAVIGATION AND OBJECT DETECTION FOR BLIND PERSON USING DEEP LEARNING

• Main Author: Lim, Yi Swen
• Format: Final Year Project Report
• Language: English; English
• Published: Universiti Malaysia Sarawak (UNIMAS) 2019
• Subjects: QA76 Computer software
• Online Access: http://ir.unimas.my/id/eprint/33719/1/Lim%20Yi%20Swen%20-%2024%20pgs.pdf; http://ir.unimas.my/id/eprint/33719/4/Lim%20Yi%20Swen%20ft.pdf; http://ir.unimas.my/id/eprint/33719/

For blind people, navigation and recognizing objects are activities which are challenging due to their disability. Therefore, their quality of life is affected to some extend and in situation where they encounter a new environment, it might contain harmful objects. The traditional methods to assist blind people to navigate such as walking stick and guide dog do provide some assistance to some extent but still does not provide accurate information to the blind person. There is also technology assisted navigation approach such using sensors and so on, but it does not provide actual information about the objects in the environment to the blind person. Therefore, this project proposed an assisted navigation and object detection solution using deep learning approach on a low-cost single board computer to allow the recognition to be done on the site itself without relying on external Internet/network connection. This project promotes a low-cost, portable, scalable and edge computing solution. It is expected that the proposed solution will enable the blind people to navigate and detect objects and improve their quality of life.