metadata
license: cc-by-4.0
language:
- ar
- fr
- en
pretty_name: semantic segmentation buildings morocco
task_categories:
- mask-generation
- image-segmentation
- feature-extraction
- image-feature-extraction
tags:
- Morocco
- Satellite
- Semantic
- Segmentation
- buildings
- mask
- SAM3
- Geospatial
+1 Million Building Segmentation Masks for Morocco
This dataset contains over 1 million building segmentation masks across Morocco, generated using the SAM3 mask generation model developed by Meta AI.
The masks were inferred from high-resolution satellite imagery (512×512 tiles at zoom level 17), covering approximately 57,000 tiles. All outputs are georeferenced and ready for download, visualization, and analysis.
Each mask includes an associated confidence score to support quality assessment and filtering for downstream tasks.
This dataset is released for open use in training, research, and exploration.