Zhenpeng Huang

VIABench: A Comprehensive Video Benchmark Collected from Blind Individuals for Visual Impairment Assistance

Yunfeng Liu, Yuandong Yang, Jiarui Han, Zhenpeng Huang, Yuqing Tang, Xiangyu Zeng, et al.

Visually impaired individuals (VIIs) encounter significant daily challenges due to limited access to visual information. Although Multimodal Large Language Models (MLLMs) have achieved impressive results on general vision and language tasks, their practical utility in real-world…Read more