# GuideDog: A Real-World Egocentric Multimodal Dataset for Blind and Low-Vision Accessibility-Aware Guidance

arXiv: https://arxiv.org/abs/2503.12844 · HTML: https://mlsanity.com/paper/2503.12844/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 17 Mar 2025.

## Built on by (most cited first, 10 of 10)

- [EgoVITA: Learning to Plan and Verify for Egocentric Video Reasoning](https://mlsanity.com/paper/2511.18242/) (2025)
- [mmWalk: Towards Multi-modal Multi-view Walking Assistance](https://mlsanity.com/paper/2510.11520/) (2025)
- [Situat3DChange: Situated 3D Change Understanding Dataset for Multimodal Large Language Model](https://mlsanity.com/paper/2510.11509/) (2025)
- [Urban Risk-Aware Navigation via VQA-Based Event Maps for People with Low Vision](https://mlsanity.com/paper/2605.11782/) (2026)
- [VIABLE: A Visually Impaired Assistance Benchmark for VLM-as-a-Judge Evaluation](https://mlsanity.com/paper/2605.31351/) (2026)
- [Believing without Seeing: Quality Scores for Contextualizing Vision-Language Model Explanations](https://mlsanity.com/paper/2509.25844/) (2025)
- [VIABench: A Comprehensive Video Benchmark Collected from Blind Individuals for Visual Impairment Assistance](https://mlsanity.com/paper/2607.14660/) (2026)
- [MAP: A Benchmark on Multimodal Accessibility Planning for Real World Places](https://mlsanity.com/paper/2608.28384/) (2026)
- [Lightweight and Resource-Efficient Perception for Robotic Guide Dogs](https://mlsanity.com/paper/2610.03187/) (2026)
- [NavAble: A Large-Scale Dataset and Synthetic Data Generation Pipeline for Blind Navigation](https://mlsanity.com/paper/2610.19330/) (2026)

## Abstract (arXiv)

For people affected by blindness and low vision (BLV), safe and independent navigation remains a major challenge, impacting over 2.2 billion individuals worldwide. Although multimodal large language models (MLLMs) offer new opportunities for assistive navigation, progress has been limited by the scarcity of accessibility-aware datasets, because creating them requires labor-intensive expert annotation. To this end, we introduce GuideDog, a novel dataset containing 22K image-description pairs (2K human-verified) capturing real-world pedestrian scenes across 46 countries. Our human-AI pipeline shifts annotation from generation to verification, grounded in established BLV guidance standards from experts and research, improving scalability while maintaining quality. We also present GuideDogQA, an 818-sample benchmark evaluating object recognition and depth perception. Experiments reveal that depth perception and adherence to these standards remain challenging for current MLLMs.