A dataset of images and corresponding semantic segmentation masks for hair, converted from COCO-style annotations. The dataset was created by author 'soychick' and last updated on September 30, 2025. It is built specifically for fine-tuning the SegFormer semantic segmentation model.
Use Cases
- Fine-tuning semantic segmentation models based on the provided image-mask pairs.
- Benchmarking hair segmentation algorithms based on the single-channel PNG mask format.
- Training models for computer vision applications involving hair detection based on the class label structure.
Strengths
- Data is formatted specifically for compatibility with the SegFormer model architecture.
- Masks are provided as single-channel PNG files with a defined ignore index (255).
Limitations
- Description metadata is limited; actual data quality requires manual inspection after download.
- Row count is unknown, which may limit suitability assessment.
- Column-level documentation is absent; field semantics must be inferred after download.
Provenance
- Source
- Generated from COCO-style annotations.
- Collection Method
- Conversion from existing annotations to a SegFormer-ready format.
- Time Range
- null
- Freshness
- Last updated 2025-09-30 05:54:25; freshness should be verified.
- Geography
- null