Assisting Independent Travel for Blind People Using Smartphone-Based Object Detection
Jee‐Eun Kim, Ayuna Okada, Akane Hoya, Masahiro Bessho · 2024
Access to environmental information, such as street signs, is crucial for blind people to travel independently. Recent advances in computer vision and machine learning techniques have opened up new opportunities to aid blind travelers by informing them of the presence of potentially significant objects in their surroundings. In this paper, we focus on the feasibility of leveraging such techniques to assist blind people in two travel tasks: operating vending machines and accessing barrier-free restrooms in public spaces. We implemented two smartphone application prototypes that provide voice guidance to aid blind travelers in performing these tasks. Specifically, one prototype helps them detect relevant objects for a vending machine (e.g., products and coin acceptors), while the other prototype helps them detect a restroom facilities (e.g., toilets and handrails). We conducted a user study on the use of our prototypes to perform these tasks with eight blindfolded participants. The evaluation results from the study demonstrate the applicability of our approach in enhancing the environmental awareness of blind travelers.