benedictpepper/Brain-Tumor-Segmentation

Brain Tumor Segmentation using Fuzzy C-Means (FCM)

An automated medical image analysis tool that applies the Fuzzy C-Means (FCM) soft clustering algorithm to segment brain MRI scans and highlight potential tumor regions.

📌 Project Overview

Diagnosing brain tumors accurately from Magnetic Resonance Imaging (MRI) is critical but often subjective and time-consuming. This project provides an automated segmentation pipeline using Fuzzy Logic. By clustering pixel data into distinct tissue groups (background, healthy tissue, fluids, and potential tumors), this tool serves as a reliable aid for medical image analysis.

🛠️ Tech Stack & Tools

• —Language: Python
• —Computer Vision: OpenCV
• —Machine Learning / Fuzzy Logic: scikit-fuzzy, NumPy
• —Visualization: Matplotlib, Graphviz (for automated process mapping)
• —GUI / Interactions: Tkinter

🚀 Key Features

• —Fuzzy C-Means Implementation: Configured to dynamically group MRI pixels into 4 distinct clusters with a fuzziness parameter of 2.0, ideal for handling the blurred boundaries of medical tissues.
• —Automated Workflow Mapping: Dynamically generates and displays Graphviz flowcharts mapping both the overall system process and the internal FCM algorithmic loops.
• —Interactive File Selection: Built-in Tkinter GUI for seamless, user-friendly local image selection and processing.
• —Quantitative Evaluation: Automatically calculates and outputs the Fuzzy Partition Coefficient (FPC) and objective function convergence to validate clustering quality.

📂 Repository Structure

fcm-brain-tumor-segmentation/
├── data/
│   └── sample_mri.jpg        # Example MRI scans for testing
├── src/
│   └── segment_fcm.py        # Main execution script
├── requirements.txt          # Python dependencies
└── README.md

💻 How to Run Locally

1. 1.Clone the repository:

   git clone https://github.com/bennypepper/Brain-Tumor-Segmentation.git
   cd brain-tumor-segmentation

1. 1.Install the required dependencies:

   pip install -r requirements.txt

Note for Windows Users: Graphviz requires a system-level installation. If the executable is missing from your PATH, the script will open an interactive prompt to help you locate the `bin` folder automatically.

1. 1.Execute the segmentation script:

   python src/segment_fcm.py

📊 Results & Performance

The algorithm efficiently reached optimal convergence in 34 iterations with a minimized objective function of 10,852,001.55.

• —Fuzzy Partition Coefficient (FPC): 0.9156 (A score highly close to 1.0, indicating exceptionally well-defined and compact clusters).
• —Visual Output: The system outputs a side-by-side comparison of the original MRI, the isolated cluster labels, and a color-mapped segmentation where potential tumor mass is isolated.

🔮 Future Development

• —Ground Truth Validation: Compare algorithmic outputs against expert-annotated masks to calculate exact Accuracy, Specificity, and Dice Coefficients.
• —Skull Stripping Pipeline: Introduce a pre-processing step to isolate brain tissue from the skull for improved clustering focus.
• —Texture Analysis Integration: Expand feature extraction beyond pixel intensity to include tissue texture matrices.