Object detection and voice alert system

In this paper we propose a method that would assist visually impaired person to navigate and avoid collision with objects, with the help of mobile application. The proposed method provides voice alert about the objects present in the path of navigation. The intent of this work is to help the differently-abled person to navigate with ease and avoid collision without being dependent on others. The object detection module of this work could be used for surveillance purpose. Object recognition is significantly influenced by issues inherent to real-world pictures, such as blurring, noise, and rotational jitters, amongst others. ML5JS, a mechanism that is based on Convolutional Neural Networks (CNN), has been used to recognize the objects in real time. As part of this work, a tracking system has been designed and tested. Person with visual disability will be able to carry the devise and avoid tripping on the objects or colliding with obstacles on the path of his navigation. The results of our work is convincing and we are confident that a bit more fine tuning of thedesign and enhancing the algorithm used will yield a devise that could be carried by the person with visual disability and navigate with confidence.