Vizwiz

The 'VizWiz' dataset is a benchmark used to evaluate vision-language models on their ability to understand and reason about complex visual scenes, particularly in the context of questions posed by visually impaired users.

Papers using Vizwiz (9)

VLM-UQBench: A Benchmark for Modality-Specific and Cross-Modality Uncertainties in Vision Language Models2026Human Uncertainty-Aware Data Selection and Automatic Labeling in Visual Question Answering2025Inverse-LLaVA: Rethinking Multimodal Alignment via Text-to-Vision Mapping2025Fast or Slow? Integrating Fast Intuition and Deliberate Thinking for Enhancing Visual Question Answering2025Believing Without Seeing: Quality Scores For Contextualizing Vision-language Model Explanations2025How to Configure Good In-Context Sequence for Visual Question Answering2023 · 3 citesOpen-Ended Visual Question Answering by Multi-Modal Domain Adaptation2019Linear Alignment of Vision-language Models for Image Captioning2023Leveraging Retrieval-Augmented Tags for Large Vision-Language Understanding in Complex Scenes2024