HPA Mask
Data set of segmentation masks generated for Human Protein Atlas cell images. Used for biomedical image classification and analysis models.
Description
The dataset HPA Mask contains over 43,000 annotation masks corresponding to individual cells, automatically generated from Human Protein Atlas images. These masks are intended to facilitate cellular classification and image segmentation in microscopy.
What is this dataset for?
- Training semantic segmentation models in a biomedical context
- Complete Human Protein Atlas images for cell classification projects
- Create research tools in AI-assisted cell biology
Can it be enriched or improved?
Yes. It is possible to combine these masks with the raw images from the Human Protein Atlas for comprehensive supervised training. Post-processing methods can also be applied to correct possible segmentation errors or create multi-layer masks for finer annotation.
🔎 In summary
🧠 Recommended for
- Bioinformatics researchers
- Microscopic segmentation projects
- Medical model training
🔧 Compatible tools
- U-Net
- MONAI
- PyTorch
- Keras
- BioImage.io
💡 Tip
Combine these masks with the original HPA images to form a complete medical segmentation pipeline.
Frequently Asked Questions
Does this dataset include the source images associated with the masks?
No, only segmented masks are included. It is recommended that they be combined with the raw Human Protein Atlas images that are available separately.
Can this dataset be used in a commercial environment?
Yes, the CC0 license allows unrestricted commercial use, including in health applications.
What is the quality of the masks provided?
The masks were generated automatically via CellSegmentator (HPA), offering robust quality but may require slight post-processing.
