Tumor Detection Model

This project is a deep learning-based tumor detection system using PyTorch and EfficientNetV2. It trains a model to classify brain MRI images as either containing a tumor ("yes") or not ("no").

Prerequisites

Before running this project, you need to have the following installed:

• Works on Macbook Sequoia 15.3.2 using an Intel Core i5, and Windows Machine with Intel Core i7

• Python 3.9.21 or 3.11.3

• PyTorch 2.1.0

• torchvision 0.16.0

• Pillow 10.0.0

• numpy 1.24.0

• Other dependencies listed in requirements.txt

You can install the required packages using pip:

pip install -r requirements.txt

Environment Setup

This project requires setting an environment variable to locate your dataset:

# Windows (Command Prompt)
set CEC_2025_dataset=path\to\your\dataset

# Windows (PowerShell)
$env:CEC_2025_dataset = "C:\full\path\to\your\dataset"

# macOS/Linux
export CEC_2025_dataset=/path/to/your/dataset

Important Notes:

1. Use the FULL path to your dataset directory
2. The path should point to the parent directory containing the yes, no, and CEC_test folders
3. In PowerShell, the environment variable will only persist for the current session

The script will use this environment variable to find your dataset directory.

Dataset Structure

The dataset should be organized as follows:

CEC_2025_dataset/
└── CEC_2025/
    ├── yes/
    │   ├── yes__001.png
    │   ├── yes__002.png
    │   └── ...
    ├── no/
    │   ├── no__001.png
    │   ├── no__002.png
    │   └── ...
    └── CEC_test/
        └── test_001.png
        └── test_002.png
        └── ...

• The yes folder contains MRI images with tumors
• The no folder contains MRI images without tumors
• The CEC_test folder contains test images for prediction

Workflow Sequence Diagram

The following sequence diagram illustrates the workflow of the tumor detection system:

sequenceDiagram
    participant User
    participant run.py
    participant Model
    participant Dataset

    User->>User: Set CEC_2025_dataset environment variable
    User->>run.py: Execute run.py
    run.py->>Model: Load trained model (final_model.pth)
    run.py->>Dataset: Load test images from CEC_test folder
    loop For each test image
        run.py->>Model: Process and predict image
        Model-->>run.py: Return prediction & confidence score
        run.py->>run.py: Classify probability (Very Unlikely to Very Likely)
        run.py->>run.py: Add result to output data
    end
    run.py->>User: Save results to output.csv
    run.py->>User: Display average confidence score

Loading

Setting up the Environment for the Model

1. Clone the repository to your local machine.

2. Open a command terminal and navigate to the python folder.

3. Install the required dependencies by running:

   pip install -r requirements.txt

4. You can now start training the model by running:

   python train.py

   Alternatively, to run the test script, execute:

   python run.py

Running the Model

To run the model on the CEC_test dataset (after setting environment var.):

python run.py

The run script will:

1. Load the trained model from final_model.pth file
2. Process all images in the CEC_test folder
3. Generate predictions with confidence scores
4. Save results to output.csv in same directory as script

Additional Scripts

Testing the Model (test.py)

To test the model's performance on the CEC_test dataset:

python test.py

The test script will:

1. Load the trained model from final_model.pth file
2. Process all images in the CEC_test folder
3. Generate predictions with confidence scores
4. Save results to output.csv

You can modify NUM_IMAGES in test.py to change the number of test images (default: 50).

Training the Model (train.py)

To train the model, run:

python train.py

The training script will:

1. Load and preprocess the images from the yes and no folders
2. Train an EfficientNetV2 model on these images
3. Save the trained model as tumor_model_1000.pth

You can modify the following parameters in train.py:

• NUM_IMAGES: Number of images to use for training
• MODEL_NAME: Name of the saved model file
• epochs: Number of training epochs

Understanding the Results

After running the script, you will see:

• A CSV file (output.csv) containing:
  • File name
  • Prediction (Yes/No)
  • Confidence score (0-1)
  • Probability classification (Very Unlikely, Unlikely, Likely, Very Likely)
• Average confidence score across all tested images
• Total number of images tested

The confidence score interpretation:

• < 0.5: Very Unlikely
• 0.5-0.75: Unlikely
• 0.75-0.9: Likely

• 0.9: Very Likely

Troubleshooting

Virtual Environment Setup

It's recommended to use a virtual environment to avoid package conflicts. Here's how to set it up:

1. Create a new virtual environment:

# Windows
python -m venv venv

# macOS/Linux
python3 -m venv venv

2. Activate the virtual environment:

# Windows (Command Prompt)
venv\Scripts\activate.bat

# Windows (PowerShell)
venv\Scripts\Activate.ps1

# macOS/Linux
source venv/bin/activate

3. Install the required packages:

pip install -r requirements.txt

4. When you're done, you can deactivate the virtual environment:

deactivate

PyTorch Installation Issues

If you encounter PyTorch compatibility issues, make sure you have Python 3.9 and PyTorch 2.1.0 installed. For newer Python versions, you may need to use the latest pre-release version of PyTorch:

pip install --pre torch torchvision torchaudio

Dataset Issues

If the scripts cannot find the dataset, verify that:

1. The environment variable CEC_2025_dataset is correctly set
2. Your folder structure matches the one described above
3. The images in the CEC_test folder are in a supported format (png, jpg, jpeg)