Vinmap: Vietnam Geographic Maps for Omission Detection and OCR

VinAIResearch released this dataset in 2024 to support computer vision research into detecting omissions in geographic maps. It focuses on Vietnamese cartography and incorporates Optical Character Recognition (OCR) elements for map analysis as presented in the MAPR'24 publication. The data enables the identification of missing geographic features or labels through automated visual inspection.

Use Cases

• Training computer vision models to detect missing labels or features in Vietnamese maps
• Developing OCR systems specialized for cartographic text extraction
• Benchmarking map omission detection algorithms as described in the MAPR'24 paper

Strengths

• Focuses on Vietnam-specific geographic maps
• Associated with peer-reviewed research from MAPR'24
• Includes OCR-related features for map text analysis

Limitations

• Unknown record count and total dataset size
• Geographic bias toward Vietnam limits global generalizability
• Lack of detailed schema documentation for annotation fields

Provenance

Source

VinAIResearch, MAPR'24 publication

Collection Method

annotated

Freshness

Last updated July 2024; reflects map data relevant to the MAPR'24 research cycle.

Geography

Vietnam