Research map / Vision-language and multimodal / Visual question answering
VizWiz Grand Challenge: Answering Visual Questions From Blind People
Danna Gurari, Qing Li, Abigale J. Stangl, Anhong Guo, Chi Lin, Kristen Grauman, Jiebo Luo, Jeffrey P. Bigham
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Fine-tune VQA V2.0 models on VizWiz | Q+I, Q+I+A, and Q+I+BUA · VizWiz test | Accuracy: 0.137 -> 0.466; 0.145 -> 0.469; 0.134 -> 0.475 | Improved |
| Train the models from scratch on VizWiz | Q+I, Q+I+A, and Q+I+BUA · VizWiz test | Accuracy: 0.137 -> 0.465; 0.145 -> 0.469; 0.134 -> 0.469 | Improved |
| Use VizWiz training rather than fine-tuning | Q+I, Q+I+A, and Q+I+BUA · VizWiz test | Accuracy: 0.466 vs 0.465; 0.469 vs 0.469; 0.475 vs 0.469 | No clear effect |
| Add attention to image and question information | Q+I · VizWiz test | Accuracy: 0.137 -> 0.145 | Improved |
| Add image features to question features | Answerability classifier · VizWiz test answerability | Average precision: 0.490 -> 0.717 | Improved |
| Add question information to image features | Answerability classifier · VizWiz test answerability | Average precision: 0.640 -> 0.717 | Improved |
Closest papers by meaning
- Grounding Answers for Visual Questions Asked by Visually Impaired People (CVPR 2022 · similarity 0.92)
- VQA: Visual Question Answering (IJCV 2016 · similarity 0.92)
- Towards VQA Models That Can Read (CVPR 2019 · similarity 0.91)
- Answer Them All! Toward Universal Visual Question Answering Models (CVPR 2019 · similarity 0.91)
- When Open-Vocabulary Visual Question Answering Meets Causal Adapter: Benchmark and Approach (AAAI 2025 · similarity 0.89)
- Right this way: Can VLMs Guide Us to See More to Answer Questions? (NeurIPS 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
