KidneyXpert

Overview

KidneyXpert is an AI-powered kidney disease classification system utilizing deep learning for automated diagnosis. This project integrates MLflow for experiment tracking and DVC for version control, providing a robust pipeline to evaluate kidney health and detect potential conditions.

Features

• AI-based kidney disease classification.
• Deep learning model trained using TensorFlow/Keras.
• Flask API for easy interaction.
• Experiment tracking with MLflow.
• Data version control using DVC.

Installation and Setup

Step 1: Clone the Repository

git clone https://github.com/danula-rathnayaka/KidneyXpert.git
cd KidneyXpert

Step 2: Create a Conda Environment

conda create -n cnncls python=3.8 -y
conda activate cnncls

Step 3: Install Dependencies

pip install -r requirements.txt

Running the Application

Before running the application, ensure that the model is built using the /train endpoint.

Step 4: Train the Model

curl -X POST http://localhost:5000/train

Step 5: Start the Application

python app.py

Now, open your browser and navigate to http://localhost:5000 to access the application.

API Endpoints

• Home Page: GET / - Serves the frontend interface.
• Train Model: POST /train - Triggers model training using DVC.
• Predict Disease: POST /predict - Accepts an image input and returns the classification result.

Model Training Pipeline

KidneyXpert follows a structured pipeline for data processing and model training:

1. Data Ingestion: Fetch and preprocess data.
2. Prepare Base Model: Create the deep learning architecture.
3. Train the Model: Train the model using TensorFlow/Keras.
4. Evaluate the Model: Use MLflow to track performance.

---