GuideDog: A Real-World Egocentric Multimodal Dataset for Blind and Low-Vision Accessibility-Aware Guidance
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913074042109952 |
|---|---|
| author | Kim, Junhyeok Park, Jaewoo Park, Junhee Lee, Sangeyl Chung, Jiwan Kim, Jisung Joung, Ji Hoon Yu, Youngjae |
| author_facet | Kim, Junhyeok Park, Jaewoo Park, Junhee Lee, Sangeyl Chung, Jiwan Kim, Jisung Joung, Ji Hoon Yu, Youngjae |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_12844 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | GuideDog: A Real-World Egocentric Multimodal Dataset for Blind and Low-Vision Accessibility-Aware Guidance Kim, Junhyeok Park, Jaewoo Park, Junhee Lee, Sangeyl Chung, Jiwan Kim, Jisung Joung, Ji Hoon Yu, Youngjae Computer Vision and Pattern Recognition 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. |
| title | GuideDog: A Real-World Egocentric Multimodal Dataset for Blind and Low-Vision Accessibility-Aware Guidance |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.12844 |