Long-Form Answers to Visual Questions from Blind and Low Vision People
This work addresses accessibility for BLV people by improving long-form visual question answering, though it is incremental as it builds on existing VQA models and datasets.
The authors tackled the problem of generating long-form visual question answers (LFVQA) for blind and low vision (BLV) users by introducing VizWiz-LF, a dataset with 4.2k answers to 600 questions, and found that while BLV users perceive both human and generated answers as plausible, generated ones often hallucinate incorrect details, especially for unanswerable questions.
Vision language models can now generate long-form answers to questions about images - long-form visual question answers (LFVQA). We contribute VizWiz-LF, a dataset of long-form answers to visual questions posed by blind and low vision (BLV) users. VizWiz-LF contains 4.2k long-form answers to 600 visual questions, collected from human expert describers and six VQA models. We develop and annotate functional roles of sentences of LFVQA and demonstrate that long-form answers contain information beyond the question answer such as explanations and suggestions. We further conduct automatic and human evaluations with BLV and sighted people to evaluate long-form answers. BLV people perceive both human-written and generated long-form answers to be plausible, but generated answers often hallucinate incorrect visual details, especially for unanswerable visual questions (e.g., blurry or irrelevant images). To reduce hallucinations, we evaluate the ability of VQA models to abstain from answering unanswerable questions across multiple prompting strategies.