Characteristics and challenges of object search in three-dimensional space via smartphone applications for visually impaired users

In recent years, the development of applications that utilize image recognition and artificial intelligence to read text and describe the surrounding environment for people with visual impairment has increased. This study investigates the actual use of these applications in three-dimensional space and the processes that arise when users search for objects. Eight participants were instructed to perform tasks simulating long-, medium-, and short-range object-search scenarios and were subsequently interviewed. A qualitative analysis revealed that the participants narrowed down spatial information step-by-step by integrating application feedback with vision, touch, and prior experience to estimate object location and proceed with exploration. This strategy is referred to as ATARI in Japanese, which predicts an unknown situation. Additionally, differences based on varying visual functions emerged among the participants. Individuals with low vision, who were able to use the gaze system to locate task-relevant objects, exhibited behaviors different from those with central scotoma, light perception, or blindness. Individuals with low and relatively preserved vision may require higher-resolution visual information than those with more severe visual impairment. These findings suggest the importance of developing customizable application interfaces customized to diverse visual abilities as well as designing physical environments that support exploratory behavior.