Kidney Disease Classifier

Overview

The Kidney Disease Classifier is a web application designed to analyze CT scan images and predict whether the scan is normal or indicates the presence of a tumor. Leveraging the VGG16 architecture as a base model, the classifier has been fine-tuned with a dataset of CT scan images, achieving an accuracy of 89%. The model is modular, allowing for enhanced accuracy with an increased number of epochs during training.

Features

• Image Upload: Users can upload CT scan images for analysis.
• Prediction Result: The application displays predictions, changing the result box color to green for normal scans and red for scans with tumors.
• Modular Design: The architecture is designed for easy scalability and adjustments.

Tech Stack

• Deep Learning Framework:
  • TensorFlow (for model training and inference)
• Data Handling:
  • Pandas (for data manipulation)
  • NumPy (for numerical operations)
• Model Management:
  • MLflow (for tracking experiments and managing models)
  • DVC (Data Version Control for dataset versioning)
• Visualization:
  • Matplotlib (for plotting)
  • Seaborn (for statistical data visualization)
• Web Framework:
  • Flask (for building the web application)
  • Flask-CORS (to handle Cross-Origin Resource Sharing)
• Other Libraries:
  • Python-box (for managing configurations)
  • PyYAML (for YAML parsing)
  • TQDM (for progress bars in loops)
  • Ensure (for ensuring type hints)
  • Joblib (for model serialization)
  • Types-PyYAML (for type checking)
  • SciPy (for scientific computations)
  • Gdown (for downloading files from Google Drive)

Installation

To set up the project locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/mShubham18/kidney-disease-dl-project.git
   cd kidney-disease-dl-project

2. Install the required packages:

   pip install -r requirements.txt

3. Run the application:

   python app.py

Usage

1. Open the web application in your browser.
2. Upload a CT scan image using the provided upload button.
3. Click on the 'Predict' button to receive the analysis result.

Contributing

Contributions are welcome! If you have suggestions or improvements, please open an issue or submit a pull request.

Acknowledgments

• VGG16 Model
• MLflow
• DVC

Happy Coding ;)