@Bench: Benchmarking Vision-Language Models for Human-centered Assistive Technology
This work addresses the problem of benchmarking vision-language models for human-centered assistive technologies, specifically for people with visual impairments, representing a novel contribution in this domain.
The paper tackles the lack of benchmarks for vision-language models in assistive technologies for people with visual impairments by creating @Bench, a novel benchmark covering five crucial tasks, and proposes @Model, which addresses all tasks simultaneously with outstanding performance.
As Vision-Language Models (VLMs) advance, human-centered Assistive Technologies (ATs) for helping People with Visual Impairments (PVIs) are evolving into generalists, capable of performing multiple tasks simultaneously. However, benchmarking VLMs for ATs remains under-explored. To bridge this gap, we first create a novel AT benchmark (@Bench). Guided by a pre-design user study with PVIs, our benchmark includes the five most crucial vision-language tasks: Panoptic Segmentation, Depth Estimation, Optical Character Recognition (OCR), Image Captioning, and Visual Question Answering (VQA). Besides, we propose a novel AT model (@Model) that addresses all tasks simultaneously and can be expanded to more assistive functions for helping PVIs. Our framework exhibits outstanding performance across tasks by integrating multi-modal information, and it offers PVIs a more comprehensive assistance. Extensive experiments prove the effectiveness and generalizability of our framework.