🧠 Stroke CT Scan Preprocessing

This repository contains utility scripts and preprocessing workflows for the TEKNOFEST 2021 stroke detection dataset.
Both DICOM and PNG brain CT scan images are processed and labeled for use in AI model training.

---

📁 Repository Structure

stroke-preprocessing/
├── utils/
│   ├── DicomHandler.py          # Handles DICOM reading & preprocessing
│   └── png_utils.py             # Handles PNG reading & normalization
│
├── notebooks/
│   └── preprocessing.ipynb      # Main Colab notebook: downloads + processes data
│
├── README.md
└── .gitignore

---

🚀 Get Started in Colab

---

🧪 Features

• ✅ DICOM & PNG image parsing
• ✅ Intensity windowing & rescaling
• ✅ Automatic resizing to 224x224
• ✅ Label assignment (0 = No stroke, 1 = Stroke)
• ✅ Train/test dataset preparation

---

📦 Requirements (handled in notebook)

• pydicom, pylibjpeg, opencv-python, numpy, matplotlib, scikit-learn

---

📊 Dataset Source

Kaggle: İNME VERİ SETİ (STROKE DATASET)
Licensed for research use under TEKNOFEST 2021 AI Competition guidelines.

---

📬 Contact

Created by Skolyn Team – Contributions welcome!