This project build a VD pipeline to describe a scene (from an image) with the hope that it could guide visually impaired people. To this end, we sequentially use a high level image-to-text model ...