Object Recognition App for Visually Impaired
Sumitra A. Jakhete, Pranit Bagmar, Avanti Dorle, Atharva Rajurkar, Piyush Pimplikar · 2019
Vision is one of the most important senses that help people interact with the real world. There are nearly 200 million blind people all over the world, and being visually impaired hinders a lot of day to day activities. Thus it is very necessary for blind people to understand their surroundings, and to know what objects they interact with. This project proposes an android application to help blind people see through handheld device like mobile phone. It integrates various techniques to build a rich android application that will not only recognize objects around visually impaired people in real time but also give an audio output to assist them as quickly as possible. SSD (Single Shot Detector) Algorithm is used for the object recognition as well as detection. Also this algorithm gives nearly accurate results for real time object detection and is proven to be faster than other relative algorithms. The application further uses android tensorflow APIs and android TextToSpeech API to give audio output.