🧠 Brain Tumor Classifier (CT + MRI)
This model uses a multimodal DenseNet architecture to classify brain tumors based on CT and MRI scans. It was trained on paired data with class labels: Healthy or Tumour.
🔬 Model Details
- Architecture: Dual DenseNet201 backbones + fusion classifier
- Input modalities: CT image, MRI image (either or both)
- Output: Binary classification (
Healthy,Tumour) - Framework: PyTorch
📦 Files
multimodal_brain_tumor_model.pth: Pretrained weights- Intended to be used with a Streamlit app (see GitHub Repo).
🚀 Usage
from huggingface_hub import hf_hub_download
import torch
path = hf_hub_download("lukmanaj/brain-tumor-multimodal", "multimodal_brain_tumor_model.pth")
model.load_state_dict(torch.load(path))
📊 Training Performance
| Epoch | Train Loss | Accuracy |
|---|---|---|
| 1 | 0.1552 | 94.82% |
| 5 | 0.0368 | 98.78% |
Note: The model shows signs of overfitting; further validation and augmentation are advised.
🧑⚕️ Intended Use
- Designed for educational and research purposes.
- Not certified for clinical or diagnostic use.
📚 Citation
Aliyu, L. (2025). Brain Tumor Classification using Multimodal Deep Learning
🤝 Acknowledgements
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support