A Fast and Comprehensive Indoor Scene Understanding Approach for Blind People

Scene perception and understanding is regarded as a major prospect for blind and visually impaired people. To date, there is no system than can completely recover the sight to them. However, various designs intended to improve their quality of life have been proposed over the last years, and have proven that assistive technologies can provide a significant help towards achieving such a prospect. The vast majority of existing assistive prototypes tends to tackle the blind rehabilitation issue in two ways. The first one includes all sorts of navigational aids, while the second one refers to object recognition. In this context, we intend in this paper to contribute with novel assistive computer-vision-based designs, to solve the problem of fast and comprehensive indoor scene description for the blind. Proposed strategies have been conducted on real images and have proven to be prominent regarding obtained results.