Detection and Classification of Kidney Diseases Using CT-Scanned Images

Project Description

This project aims to detect and classify kidney diseases (Cyst, Normal, Stone, Tumor) from CT-scanned images. Leveraging machine learning and deep learning techniques, the project automates the classification process to assist medical professionals in diagnostics.

Dataset

• Dataset Name: CT-KIDNEY-DATASET
• Categories:
  1. Cyst
  2. Normal
  3. Stone
  4. Tumor
• Structure:
  The dataset is organized into subdirectories for each class. Each subdirectory contains CT scan images corresponding to its label.

Requirements

Python Libraries

• numpy
• matplotlib
• seaborn
• tensorflow
• keras
• scikit-learn
• imblearn (for SMOTE)
• Pillow (for image processing)
• os and shutil (for file operations)

Install all dependencies with:

pip install -r requirements.txt  

Folder Structure

project_root/  
│  
├── CT-KIDNEY-DATASET-Normal-Cyst-Tumor-Stone/  
│   ├── Cyst/  
│   ├── Normal/  
│   ├── Stone/  
│   └── Tumor/  
│  
├── notebooks/  
│   ├── data_visualization.ipynb  
│   ├── model_training.ipynb  
│   └── evaluation.ipynb  
│  
├── models/  
│   └── saved_model.h5  
│  
├── README.md  
├── requirements.txt  
└── main.py  

Workflow

1. Data Preprocessing

• Image Resizing: All images are resized to a consistent shape (e.g., 32x32).
• Normalization: Pixel values are normalized to a range of [0, 1].
• Class Balancing: SMOTE (Synthetic Minority Oversampling Technique) is used to balance the dataset.

2. Data Visualization

• Distribution of classes.
• Examples of CT-scanned images for each category.

3. Model Training

• Model Architecture: Convolutional Neural Networks (CNN).
• Variants:
  • Base Model
  • Dropout for regularization.
  • Hyperparameter-tuned models.
• Optimizer: Adam.
• Loss Function: Categorical Cross-Entropy.

4. Evaluation

• Accuracy and F1 Score.
• Confusion Matrix to visualize class-wise performance.

How to Run

1. Clone this repository:

   git clone https://github.com/Oulakbir/Kidney_Deseases_Detection.git  

2. Navigate to the project directory:

   cd Kidney_Deseases_Detection  

3. Install dependencies:

   pip install -r requirements.txt  

4. Run the main script for training:

   python main.py  

5. View results and evaluation in the evaluation.ipynb notebook.

Results

Model	Accuracy	F1 Score
Base CNN	85%	0.84
CNN + Dropout	88%	0.87
Tuned CNN	91%	0.90

Or Run the app.py to use the streamlit application:

Future Work

• Explore advanced architectures like ResNet or EfficientNet.
• Extend the dataset with additional cases.
• Implement real-time prediction for CT-scanned images.

Contributors

• Your Name: LinkedIn | GitHub

License

This project is licensed under the MIT License. See the LICENSE file for more details.