Dataset · ML2
Test sets for navigation text recognition
ML2/navocr-testsets
1,000 to 10,000 labeled photos of Korean and English signs to evaluate mobile robot OCR systems.
What's inside
A NeurIPS 2026 paper dataset containing 1,000 to 10,000 annotated images for object detection and text recognition in Korean and English. The annotations include only navigationally significant text (store and place names) used by humans or robots for positioning, while menus, prices, and slogans are filtered out.
What you can do with it
- Testing OCR and text detection models for mobile robot navigation
- Benchmarking visual orientation systems in urban environments
- Tasks
- Object detection, Image → text
- Languages
- Korean, English
- Size
- 1K–10K rows
- Downloads this month
- 1.3K
- Likes
- 2
How to load it
from datasets import load_dataset
ds = load_dataset("ML2/navocr-testsets")