How manasisbyali234/brain_tumor works
This repository is a CT–MRI brain-tumor detection application with training, evaluation, inference, and visual explanation workflows. Users can submit scan pairs through the Streamlit or Tkinter interface, or run command-line prediction: images are prepared, aligned, fused, classified, and presented with explanation outputs. The graph also preserves the documented data-training and evaluation branches. Some UI and supporting source excerpts are partial; edges reflect explicit README statements, shown calls, or SOURCE INDEX references rather than assumed wiring.
Main entry points: predict.py Pair prediction · app.py Web interface · gui_app.py Desktop interface · dataset.py Dataset builder
Components 16
User interfaces
- Web interface
app.py - Desktop interface
gui_app.py
Image processing
- Scan dataset
data - Dataset builder
utils/dataset.py - Preprocessing
utils/preprocessing.py - Image registration
utils/registration.py - Image fusion
utils/fusion.py
Model and learning
- Dataset training
train.py - Image-based training
train_on_images.py - CNN models
models/cnn_model.py - Model evaluation
evaluate.py
Interpretation and results
- Pair prediction
predict.py - Explainability
utils/explainability.py - Tumor marking
mark_tumors.py - Saved results
outputs
Other components
- Clinician / user
Connections 28
- Clinician / user → Web interface: uses
README.md - Clinician / user → Desktop interface: uses
README.md - Web interface → Preprocessing: runs
app.py - Web interface → Image registration: runs
app.py - Web interface → Image fusion: runs
app.py - Web interface → Explainability: runs
app.py - Web interface → CNN models: uses
app.py - Desktop interface → Preprocessing: calls
gui_app.py - Desktop interface → Image registration: calls
gui_app.py - Desktop interface → Image fusion: calls
gui_app.py - Desktop interface → Explainability: calls
gui_app.py - Scan dataset → Dataset builder: supplies scans
README.md - Dataset builder → Preprocessing: calls
utils/dataset.py - Dataset builder → Image registration: calls
utils/dataset.py - Dataset builder → Image fusion: calls
utils/dataset.py - Dataset training → Dataset builder: loads data
train.py - Dataset training → CNN models: builds model
train.py - Image-based training → CNN models: trains
train_on_images.py - Pair prediction → Preprocessing: calls
predict.py - Pair prediction → Image registration: calls
predict.py - Pair prediction → Image fusion: calls
predict.py - Pair prediction → CNN models: predicts with
README.md - Pair prediction → Explainability: requests explanations
README.md - Pair prediction → Saved results: saves figure
README.md - Model evaluation → Saved results: saves metrics
README.md - Tumor marking → Preprocessing: calls
mark_tumors.py - Tumor marking → CNN models: loads weights
mark_tumors.py - Explainability → Saved results: optional SHAP plot
README.md