ARTIFICIAL VISION FOR THE VISUALLY IMPAIRED
Siddhi Potdar, Sakshi Gupta Siddharth Kulkarni Aditya Kawale Pallavi Dhade · Journal of Critical Reviews · 2020
Visual Question Answering (VQA) is a computer vision task where a system is given a textual question about a visual (image) input, and it generates a relevant answer. This approach can be used to devise a system to aid the visually impaired by assisting them in their day-to-day visual tasks. In order to do so the focus areas involved are: 1) Selection of an adequate dataset 2) Promising speech interpretation 3) Effective Object detection 4) Subsume appropriate methodology. We present a survey revolving around the same covering various aspects viz. types, algorithms, techniques and procedure followed in respective areas. Furthermore, we compare them and draw conclusions specific to developing a system for visually impaired.