How 141002/brain-tumor-mri-tda-ml works
This is a notebook-led brain MRI classification experiment using persistent homology. The documented workflow loads MRI images, prepares them for cubical-complex filtration, derives H0/H1 persistence-image features, then splits the data, evaluates three classifiers, and compares results. The repository tree contains only the notebook and requirements file; the README describes additional source modules, but they are not present in the supplied tree. The graph therefore maps documented responsibilities to the notebook and cites README relationships; notebook internals were not sampled.
Main entry points: brain_tumor_mri.ipynb Stratified split
Components 16
Image Preparation
- Image preparation
brain_tumor_mri.ipynb - Filtration grid
brain_tumor_mri.ipynb
Topological Features
- Cubical complex
brain_tumor_mri.ipynb - Persistence computation
brain_tumor_mri.ipynb - H0 diagrams
brain_tumor_mri.ipynb - H1 diagrams
brain_tumor_mri.ipynb - Persistence images
brain_tumor_mri.ipynb - 800D features
brain_tumor_mri.ipynb
Classification
- Stratified split
brain_tumor_mri.ipynb - Logistic regression
brain_tumor_mri.ipynb - RBF SVM
brain_tumor_mri.ipynb - Random forest
brain_tumor_mri.ipynb - Model evaluation
brain_tumor_mri.ipynb
Other components
- Notebook user
- MRI dataset
- Experiment notebook
brain_tumor_mri.ipynb
Connections 18
- Notebook user → Experiment notebook: runs
README.md - MRI dataset → Experiment notebook: supplies images
README.md - Experiment notebook → Image preparation: prepares images
README.md - Image preparation → Filtration grid: normalizes and inverts
README.md - Filtration grid → Cubical complex: builds grid
README.md - Cubical complex → Persistence computation: computes intervals
README.md - Persistence computation → H0 diagrams: produces H0
README.md - Persistence computation → H1 diagrams: produces H1
README.md - H0 diagrams → Persistence images: vectorizes
README.md - H1 diagrams → Persistence images: vectorizes
README.md - Persistence images → 800D features: concatenates
README.md - 800D features → Stratified split: feeds
README.md - Stratified split → Logistic regression: trains and tests
README.md - Stratified split → RBF SVM: trains and tests
README.md - Stratified split → Random forest: trains and tests
README.md - Logistic regression → Model evaluation: reports metrics
README.md - RBF SVM → Model evaluation: reports metrics
README.md - Random forest → Model evaluation: reports metrics
README.md