Perception transformation for visually impaired people

Abstract

Blindness is the form of absence of visual sensation due to physiological or neurological factors. With this lack of vision blind people have to face unexpected situations while travel or attend their day to day work. There are many guidance systems with a few superfluous features for visually impaired travelers to navigate quickly and safely against obstacles. Most of such systems support only one area of interest such as identifying objects and supporting navigation. The system proposed in this paper is capable of providing some essential guidance for a blind person to find the way in an in-house environment. The research focuses on exploiting the use of image processing and other sensors to provide a clear understanding of an environment with voice perception. The visual perception is captured by the camera attached to the blind person and provides preprocessed instances of these visual frames as descriptors to an artificial neural network. In addition to object recognition, sensor based hardware module is used to detect obstacles, and to provide information on other unexpected situations.