Design and Implementation of the Smart Glove to Aid the Visually Impaired

Locating objects of daily use is a strenuous task for the visually impaired. The objective of this paper is to design a smart glove by using Deep Neural Networks (DNN) and object tracking algorithm which will guide the hand of the visually impaired to the desired object in an indoor environment. The smart glove has five micro-vibrating motors, each one used to guide the user’s hand in five different directions namely, forward, upward, downward, rightward and leftward. The palm of the glove has a Universal Serial Bus (USB) camera which feeds the real-time video to the Raspberry Pi for processing. The camera also has an inbuilt microphone. The user vocally commands the system to identify the desired object. The camera then detects the object using DNN. Once the object is tagged, object tracking begins. Based on the relative position of the camera and the object, the micro-vibrating motors vibrate accordingly to guide the users hand in the required direction. An additional feature which is incorporated includes differentiation of similar objects by color.