SubhranshuSharma/lung-cancer-detection

LUNGAI: Lung Cancer Detection Model

Project Overview

LungAI is a deep learning project aimed at detecting and classifying lung cancer from CT scan images. The model can differentiate between cancerous and non-cancerous lung tissue, as well as classify specific types of lung cancer.

4x hackathon award winner - out of 1,500 total competitors.

Model Performance

• 98% accuracy in distinguishing between cancerous and non-cancerous cases
• 86% accuracy in differentiating between four specific types of lung conditions:
  • Adenocarcinoma: 84% F1-score
  • Large Cell Carcinoma: 82% F1-score
  • Normal (non-cancerous): 98% F1-score
  • Squamous Cell Carcinoma: 83% F1-score

This project represents the newest version, now using PyTorch.

Repository Structure

• Architecture/: Contains the core model scripts
  • architecture.py: Defines the model architecture
  • preprocess.py: Data preprocessing utilities
  • test.py: Script for testing the model
• Model/: Stores trained model files
  • lung_cancer_detection_model.onnx: ONNX format of the trained model
  • lung_cancer_detection_model.pth: PyTorch weights of the trained model
• Data/: (Not included in repository) Directory for storing the dataset
• Processed_Data/: (Not included in repository) Directory for preprocessed data
• assets/: Additional project assets
• requirements.txt: List of Python dependencies

Setup and Usage

Step 1: Install Dependencies

First, ensure you have Python installed. Then, install the required Python libraries using the following command:

pip install -r requirements.txt

Step 2: Train the Model (Optional)

Run the training script to train the model. It will be saved as .pth and .onnx files

python Architecture/architecture.py

Step 3: Run the Model

Run the model by running the following file:

python Architecture/run.py

Notes

• Make sure your dataset is structured correctly under the Processed_Data directory with subdirectories for training, validation, and testing sets.
• The model training script expects the dataset to be in the Processed_Data directory. Ensure that the data transformations and directory paths are correctly set up in architecture.py.

Contributing

If you would like to contribute to this project, please fork the repository and submit a pull request. We welcome improvements, bug fixes, and new features.

Connect with Me