News
Understanding How Blind Users Handle Object Recognition Errors
Camera-based assistive technologies like object recognizers are impactful for many people with vision impairments. The insights that authors gained from this study can be used to improve these technologies are reduce the number of errors.
UMD researchers Jonggi Hong and Hernisa Kacorri had the opportunity to present their research on the use of object recognition technology at the 2024 ASSETS conference (an accessibility and computing conference).
“Object recognition technologies hold the potential to support Blind and low-vision people in navigating the world around them. However, the gap between benchmark performances and practical usability remains a significant challenge” (Kacorri et. al, 2024).
In their study, the authors conducted a user study involving 12 blind and low-vision participants and leveraged a pre-existing object recognition system, URCam, fine-tuned for their experiment. Through in-depth interviews and hands-on error identification tasks, the authors gained insights into users’ experiences, challenges, and strategies for identifying errors in camera-based assistive technologies and object recognition systems.
Read their paper Understanding How Blind Users Handle Object Recognition Errors: Strategies and Challenges (2024).
