Cvrie

A two-part machine learning project in Python that applies AI to a real medical-imaging problem: automatically classifying X-ray scans by body part (chest, foot, hand, head). It pairs a supervised medical image classifier with an unsupervised NLP study, spanning both ends of the ML spectrum.

Features

• Supervised — medical X-ray classification — a healthcare use case: sort radiography images by body part (chest, foot, hand, head). Image loading and normalization, dataset building (Supervised/Source/), and scikit-learn models (Random Forest, MLP, SVC pipeline) with a StandardScaler; models serialized via joblib (Supervised/Notebooks/supervised.ipynb)
• Unsupervised — NLP — tokenization and text cleaning with nltk, plus clustering on a student dataset (Unsupervised/)
• Structured layout — Supervised/, Unsupervised/ and EnderChest/Bootstrap/ (warm-up notebooks), each part with its own dependencies

Build

Each part has its own requirements.txt.

python3 -m venv .venv
source .venv/bin/activate
pip install -r Supervised/requirements.txt      # or Unsupervised/requirements.txt

Usage

Open the relevant notebook in Supervised/Notebooks/ or Unsupervised/Notebooks/ (Jupyter or VS Code) and run the cells. The supervised helper scripts live in Supervised/Source/.

Trained models (Supervised/ModelSave/*.pkl) and the X-ray image dataset are not included — models are regenerated by running the notebooks. The student dataset (Unsupervised/Dataset/Student_Dataset.csv) is included.

Credits

Team project realized as part of the Epitech curriculum — Machine Learning module.

• Nielsen Combe-Bracciale
• Mathys Aberkane