Towards Independent Navigation with Visual Impairment: A Prototype of a Deep Learning and Smartphone-based Assistant

Abstract

People with visual impairments may find it hard to navigate in unfamiliar spaces. Even though several systems have been proposed to make navigation easier, such systems often ignore two essential factors for anyone using a support system: comfort and portability. This work presents a portable navigation support system for people with visual impairments. The system utilizes Smartphones as a portable platform and leverages the application of a deep learning object detection system for guiding the user during navigation. Audio output is used to communicate information about obstacles to the user. The system offers promising accuracy in object detection and distance estimation.