🧠 Brain Tumor Segmentation

📘 Project Overview

This project is dedicated to the development of an algorithm and software for brain tumor segmentation from MRI images using deep learning methods. The goal is to improve the speed and accuracy of diagnosis, which is critical for timely medical treatment.

Automation allows to:

• reduce the workload on radiologists,
• minimize the risk of human errors,
• ensure consistent image analysis quality.

---

🧩 Technologies Used

• Programming language: Python

• Libraries and frameworks:

  • TensorFlow, Keras — building and training the neural network
  • scikit-learn — data preprocessing and evaluation
  • NumPy, Matplotlib — data manipulation and visualization
  • PyQt5 — GUI development

---

🧠 Model Architecture

The model is based on the U-Net architecture, designed specifically for biomedical image segmentation. U-Net consists of an encoder-decoder structure with skip connections, which help preserve fine details during upsampling.

---

🗂️ Dataset

Four popular public datasets were analyzed:

• BraTS
• ASNR
• TSIA
• FigShare

After comparison, BraTS was selected because it includes:

• well-annotated tumor masks,
• multiple MRI modalities (T1, T2, FLAIR, etc.),
• a large number of labeled samples for training.

---

⚙️ Data Preprocessing

1. Intensity normalization
2. Image resizing
3. Data augmentation (flipping, rotation, shifting)
4. Splitting into training, validation, and test sets

---

📈 Model Training

The model was trained using the BraTS dataset. During training, loss decreased and accuracy improved, indicating proper learning behavior.

---

🧪 Results

The model was evaluated on a held-out test set that was not used during training. The trained model successfully detected tumor regions with high accuracy.

---

💻 Graphical Interface

The desktop application was implemented using PyQt5. The user interface allows:

• uploading MRI images,
• running segmentation,
• visualizing detected tumor regions.

---

🚀 Installation and Run

Installation

git clone https://github.com/username/brain-tumor-segmentation.git
cd brain-tumor-segmentation
pip install -r requirements.txt

Running the Application

python final.py

---

📊 Usage Example

---

📚 Conclusion

As a result of this work:

• an algorithm and software for brain tumor segmentation from MRI images were developed,
• a comparative analysis of modern approaches was conducted,
• a U-Net–based neural architecture was implemented,
• a user-friendly interface was created for medical image analysis.

This project can be used in diagnostic systems and research applications for medical imaging.