1 paper · 1 filter
Junbin Xiao, Nanxin Huang, Hao Qiu +5
We present EgoBlind, the first egocentric VideoQA dataset collected from blind individuals to evaluate the assistive capabilities of contemporary multimodal large language models (…