Assisting the visually impaired using depth inference on mobile devices via stereo matching

A novel system for assisting the visually impaired, built upon the platform of a mobile device, is proposed. The system uses depth inference and a depth-to-sound mapping to inform the user of the surroundings via sound. Stereo matching is proposed as the appropriate tool for depth inference, and stereo matching approaches are assessed for their potential for real-time performance on a mobile device and their inference quality. A novel depth-to-sound mapping is proposed that can be adjusted by the user through the touch interface of the mobile device. This mapping allows the user to define the field of view and the location of inference. A smart phone “app” and demo will be produced. The system demonstrates the viability and great benefit of depth inference on mobile devices for assisting the visually impaired.