|
|
--- |
|
|
library_name: transformers |
|
|
license: cc-by-nc-4.0 |
|
|
datasets: |
|
|
- histai/SPIDER-breast |
|
|
base_model: |
|
|
- histai/hibou-L |
|
|
pipeline_tag: image-classification |
|
|
--- |
|
|
|
|
|
# SPIDER-Breast Model |
|
|
|
|
|
|
|
|
### Model Description |
|
|
|
|
|
SPIDER-breast model is a deep learning model trained for patch-level pathology classification, specifically for breast. It is part of the SPIDER dataset initiative, which provides a large, high-quality, multi-organ pathology dataset with expert-annotated labels. |
|
|
|
|
|
If you would like to support, sponsor, or obtain a commercial license for the SPIDER data and models, please contact us at [email protected]. |
|
|
|
|
|
### Model Sources |
|
|
|
|
|
- **Repository:** [https://github.com/HistAI/SPIDER](https://github.com/HistAI/SPIDER) |
|
|
- **Paper:** [SPIDER: A Comprehensive Multi-Organ Supervised Pathology Dataset and Baseline Models](https://arxiv.org/abs/2503.02876) |
|
|
|
|
|
## How to Get Started with the Model |
|
|
|
|
|
Model works with **1120×1120** patches. |
|
|
Use the following code snippet to load and use the model: |
|
|
|
|
|
```python |
|
|
from transformers import AutoModel, AutoProcessor |
|
|
from PIL import Image |
|
|
|
|
|
model = AutoModel.from_pretrained("histai/SPIDER-breast-model", trust_remote_code=True) |
|
|
processor = AutoProcessor.from_pretrained("histai/SPIDER-breast-model", trust_remote_code=True) |
|
|
|
|
|
image = Image.open("path_to_image.png") |
|
|
inputs = processor(images=image, return_tensors="pt") |
|
|
outputs = model(**inputs) |
|
|
print(outputs.predicted_class_names) |
|
|
``` |
|
|
|
|
|
|
|
|
### Training Data |
|
|
|
|
|
The model is trained on the [SPIDER-breast](https://huggingface.co/datasets/histai/SPIDER-breast) dataset, a subset of the SPIDER dataset. The dataset includes: |
|
|
|
|
|
| Class | Total Patches | |
|
|
|--------------------------------------|---------------| |
|
|
| Adenosis | 2899 | |
|
|
| Benign phyllodes tumor | 4526 | |
|
|
| Ductal carcinoma in situ (high-grade)| 5632 | |
|
|
| Ductal carcinoma in situ (low-grade) | 5017 | |
|
|
| Fat | 6286 | |
|
|
| Fibroadenoma | 5243 | |
|
|
| Fibrocystic changes | 5027 | |
|
|
| Fibrosis | 6260 | |
|
|
| Invasive non-special type carcinoma | 6142 | |
|
|
| Lipogranuloma | 4941 | |
|
|
| Lobular invasive carcinoma | 5102 | |
|
|
| Malignant phyllodes tumor | 5271 | |
|
|
| Necrosis | 5396 | |
|
|
| Normal ducts | 4891 | |
|
|
| Normal lobules | 5821 | |
|
|
| Sclerosing adenosis | 3423 | |
|
|
| Typical ductal hyperplasia | 5546 | |
|
|
| Vessels | 5469 | |
|
|
|
|
|
**Total Counts:** |
|
|
- **92,892** central patches |
|
|
- **984,924** total patches (including context patches) |
|
|
- **921** total slides used for annotation |
|
|
|
|
|
### Results |
|
|
|
|
|
| Organ | Accuracy | Precision | F1 Score | |
|
|
|---------|----------|------------|----------| |
|
|
| Breast | 0.902 | 0.896 | 0.897 | |
|
|
|
|
|
## License |
|
|
The model is licensed under **CC BY-NC 4.0** and is for **research use only**. |
|
|
|
|
|
## Citation |
|
|
If you use this model, please cite the following: |
|
|
|
|
|
```bibtex |
|
|
@misc{nechaev2025spidercomprehensivemultiorgansupervised, |
|
|
title={SPIDER: A Comprehensive Multi-Organ Supervised Pathology Dataset and Baseline Models}, |
|
|
author={Dmitry Nechaev and Alexey Pchelnikov and Ekaterina Ivanova}, |
|
|
year={2025}, |
|
|
eprint={2503.02876}, |
|
|
archivePrefix={arXiv}, |
|
|
primaryClass={eess.IV}, |
|
|
url={https://arxiv.org/abs/2503.02876}, |
|
|
} |
|
|
``` |
|
|
|
|
|
## More Information |
|
|
|
|
|
To explore other models and the SPIDER dataset you can visit the [Hugging Face HistAI page](https://huggingface.co/histai) or [GitHub](https://github.com/HistAI/SPIDER) of the project. |
|
|
|
|
|
## Contacts |
|
|
|
|
|
- **Authors:** Dmitry Nechaev, Alexey Pchelnikov, Ekaterina Ivanova |
|
|
- **Email:** [email protected], [email protected], [email protected] |
