How utkarsh0141/malaria-detection-using-machine-learning- works
This is a Flask-based malaria cell-image classifier, with scripts for preparing data, training candidate models, inspecting the dataset, and evaluating a selected model. The main user flow is image upload, preprocessing, model inference, label/confidence calculation, and HTML response. The dataset itself is not present in the tree, and the training script’s saved model names do not establish how best_malaria_model.h5 was selected; those links are left unspecified.
Main entry points: app.py Flask app · train_model.py Model training · split_dataset.py Dataset splitting · evaluation.py Model evaluation
Components 15
Web inference
- Flask app
app.py - Image preprocessing
app.py - Malaria classifier
best_malaria_model.h5 - Result page
templates/index.html
Data and training
- Dataset splitting
split_dataset.py - Dataset inspection
visualization.py - Model training
train_model.py - MobileNetV2 model
MobileNetV2_model.h5 - EfficientNetB0 model
EfficientNetB0_model.h5
Model evaluation
- Model evaluation
evaluation.py - Evaluation plots
Other components
- Image user
- Cell image dataset
- Training data
- Validation data
Connections 17
- Image user → Flask app: uploads image
- Flask app → Result page: renders page
- Flask app → Image preprocessing: prepares image
- Image preprocessing → Malaria classifier: provides tensor
- Flask app → Malaria classifier: requests prediction
- Flask app → Result page: renders result
- Cell image dataset → Dataset splitting: reads images
- Dataset splitting → Training data: copies training set
- Dataset splitting → Validation data: copies validation set
- Cell image dataset → Dataset inspection: reads samples
- Training data → Model training: loads images
- Validation data → Model training: validates models
- Model training → MobileNetV2 model: saves model
- Model training → EfficientNetB0 model: saves model
- Validation data → Model evaluation: loads images
- Model evaluation → Malaria classifier: loads and predicts
- Model evaluation → Evaluation plots: displays metrics