Evaluation of feedback mechanisms for wearable visual aids

Sensory substitution devices use other sensory modalities to communicate information typically obtained through the sense of sight. In this paper, we examine the ability of blind subjects to follow simple verbal commands that allow them to accomplish two critical tasks: (1) locating and grasping an object, and (2) navigating a complex path. A total of 5 blind subjects were enrolled in the study. The object localization and tracking system used a glasses-mounted wide-field-of-view camera, a computer vision system to track the object as the subject moved the camera, and a feedback algorithm to provide verbal cues to guide the subject to center the object in the camera field of view. Subjects could successfully locate and grasp an object 60% to 100% of the time, and improved significantly with practice. In order to test navigation and mobility, a custom Android application produced verbal commands to allow an investigator to guide a given subject through a course. Subjects could follow commands with 69% to 95% compliance and reaction times of 1 to 2 seconds, and judged the system usability to be between 67.5% and 100%.