Image Recognition based SMART AID FOR VISUALLY CHALLENGED PEOPLE

摘要

This project tries to transform the visual world into the audio world with the potential to inform visually challenged people the spatial locations of the objects in their vicinity. Our paper presents the development of a real-time system based on detection, classification, and position estimation of objects in an outdoor environment to provide the visually impaired individuals with a voice output-based scene perception. The system is low-cost, light weight, simple, and easily wearable. The module is integrated into the stick, and the pi-camera is used to take the picture, and a controller is provided to move the camera in the required direction. The valuable insights gained from the feedback are then used to modify the system to best suit the requirements of the user. The object detection and classification framework exploit a multi-modal fusion-based mask RCNN using motion, sharpening and blurring filters for efficient feature representation. The image recognition classifies the detected objects along with the positions of the objects. Experimental results carried out in the outdoor environment are demonstrated. These data obtained is then converted to voice, and blind people will get the visual of what the environment is around them.