Hybrid Image Processing Device as Wearable Aide for Visually Impaired

With the advent of technology and algorithms in image processing, many wearable aides are available in the market, and researchers across the globe are developing new solutions. The existing solutions and products fail to provide an economically viable and single solution to the problem of visually impaired people. This study introduces AIde, a novel system amalgamating optimized computer vision algorithms and wearable IoT devices to convert the information extracted from the camera mounted on the user to audio-based signals for assisted mobility. The proposed system has three user modes: indoor, outdoor, and hybrid, which can be modified using the buttons provided on the wearable device. The indoor mode can be used for object detection and classification to increase the accessibility of the objects inside a confined space, and the outdoor mode will provide the user with enhanced mobility and directional sense based on the highlighted path captured by the mounted camera. The hybrid model combines both the features of indoor and outdoor modes, which will detect objects and improve mobility simultaneously. The proposed model is optimized to work in a computationally efficient environment with an at par efficiency and accuracy as the state-of-art and costly market alternatives.