Reproduce “MProtoNet: A Case-Based Interpretable Model for Brain Tumor Classification with 3D Multi-parametric Magnetic Resonance Imaging” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yuanyuan Wei, Roger Tam, Xiaoying Tang
The MProtoNet model, developed by researchers at SUSTech and UBC, introduces a case-based interpretable framework for 3D multi-parametric MRI brain tumor classification. It demonstrates enhanced interpretability through improved localization coherence (AP of 0.713) and correctness (IDS of 0.079) without requiring pixel-level annotations for training, matching the diagnostic performance of conventional CNNs.
Key insights
- The combination of Soft Masking and Online-CAM Loss synergistically enhances the correctness and localization coherence of attention maps within 3D medical data.
- Case-based interpretable models can be effectively adapted for complex 3D multi-parametric MRI data, offering intuitive explanations for brain tumor classification.
- It is feasible to improve model interpretability in medical AI without compromising diagnostic performance or relying on fine-grained, human-annotated segmentation masks during training.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.