A Review on Vision-Based Obstacle Avoidance and Assistant Systems for Visually Impaired People.
Keywords:Computer Vision, Deep Learning, Smart Wearables
Travelling is one of the biggest problems faced by the visually impaired community in their day today life. Even though the visually impaired people have their own methods for travelling like white canes and guide dogs, those traditional methods have lot of drawbacks and issues. Because of those issues, researches has paid attention to develop assistive technologies which can help the visually impaired community for their day today travelling. Although the assistant technologies are developed in the past decades, the visually impaired community face lot of issues when using them because of the Disability Digital Devide. As a result of this, number of systems and technologies have been introduced to the word. Those systems use different technologies like vision modules, sensors and GIS maps but most of those systems are vision-based and they use different approaches to develop their vision modules such as deep learning, reinforcement learning, stereo imaging, enhanced image processing techniques, etc. So in this paper, we will focus on those vision-based systems to find the optimal way of developing a vison-based assistant system for visually impaired people by studying those technologies and by comparing and contrasting those technologies by understanding their used modules algorithms, efficiency, usability, functionalities and their advantages and disadvantages.