This project focuses on Kidney Tumor Detection and Classification, integrating Machine Learning Operations (MLOps) practices and Data Version Control (DVC) for efficient project management and versioning. The core of the solution is built around the VGG16 Keras model. A Flask application serves as the interface for making predictions, providing users with an accessible and practical tool for identifying kidney tumors. The model achieves an accuracy of 84.45 percent, demonstrating its reliability and can further be increased by training it on more epochs.
Source dataset was collected from PACS (Picture archiving and communication system) from different hospitals in Dhaka, Bangladesh where patients were already diagnosed with having a kidney tumor, cyst, normal or stone findings. Both the Coronal and Axial cuts were selected from both contrast and non-contrast studies with protocol for the whole abdomen and urogram.
After downloading the dataset, zip it and push it to google drive and use gdown to download it into your system while running to save space.
- View my research folder to know more about the experiments.
To run this project :
-
git clone https://github.com/Shreyansh-1998/MLOps_CNNClassifier.git -
conda create -n your_env python=3.8 -y -
conda activate your_env/ -
pip install -r requirements.txt -
python3 main.py -
python3 app.py
This project also contains Data Version Control(Dvc) for pipeline tracking to save computational power. It is integrated with Mlflow to track experiments using Dagshub.
- Login to dagshub.
- Connect the git repository you want to track.
- Export env_variables for remote server
export [MLFLOW_TRACKING_URI]="remote_uri"
export [MLFLOW_TRACKING_USERNAME]= "your_username"
export [MLFLOW_TRACKING_PASSWORD]= "your_password"
- Run the experiments and it logs metrics and parameters
MLflow tutorial and documentation
Steps for dvc execution:
dvc init
dvc dag
dvc repro
