Best Paper Finalist at UR 2024

“Is It Safe to Cross? Interpretable Risk Assessment with GPT-4V for Safety-Aware Street Crossing” was named a Best Paper Finalist at UR 2024. Part of the guide-dog robot project, the work uses a vision-language model to judge whether a street crossing is safe and to explain its reasoning.

Led by Hochul Hwang, with co-authors Sunjae Kwon and Yekyung Kim. See the publications page for the paper.