🩺 Disease Detection and Diagnosis using Computer Vision

A deep learning-based healthcare application that leverages YOLO and Computer Vision techniques to detect and classify chest-related diseases (e.g., pneumonia, tuberculosis) from X-ray images. Built using Python, OpenCV, YOLOv5, and deployed via a Streamlit web app.

📌 Project Overview

• ✅ Detect chest anomalies using YOLOv5
• ✅ Classify diseases with CNNs or pretrained models
• ✅ Streamlit-based interactive web interface
• ✅ Visualization with bounding boxes and confidence scores
• ✅ Real-time predictions and batch processing support

🔍 Use Case

Early detection of chest diseases such as:

• Pneumonia
• Tuberculosis
• Cardiomegaly
• COVID-19 (optional)
• Normal (no abnormality)

📂 Project Structure

disease-detection-cv/
├── data/
├── models/
├── notebooks/
├── src/
│   ├── detector.py
│   ├── classifier.py
│   ├── preprocessing.py
│   ├── utils.py
├── streamlit_app.py
├── requirements.txt
├── README.md
└── yolov5/

⚙️ Features

• 🔬 YOLOv5 Detection
• 🧠 Classification
• 🌐 Web App
• 📊 Performance Metrics
• 📸 Visualization

🛠️ Installation

git clone https://github.com/yourusername/disease-detection-cv.git
cd disease-detection-cv
pip install -r requirements.txt

Download trained YOLOv5 weights and place them in the models/ directory.

🚀 Run the App

streamlit run streamlit_app.py

Upload a chest X-ray image and get disease predictions with bounding boxes and confidence scores.

📊 Model Training (Optional)

1. Label dataset using Roboflow or LabelImg
2. Convert to YOLO format
3. Train YOLOv5:

   python yolov5/train.py --img 640 --batch 16 --epochs 100 --data data.yaml --weights yolov5s.pt

🧪 Evaluation

• YOLOv5 Metrics: [email protected], Precision, Recall
• Classification Metrics: Accuracy, ROC-AUC, Confusion Matrix
• Visual Results: Bounding boxes on test images

📈 Results

Disease	Precision	Recall	F1 Score
Pneumonia	0.92	0.89	0.90
Tuberculosis	0.88	0.87	0.87
Cardiomegaly	0.90	0.91	0.90
Normal	0.95	0.96	0.95

🧰 Tech Stack

• YOLOv5
• OpenCV
• PyTorch
• TensorFlow/Keras (optional)
• Streamlit
• Matplotlib, Seaborn
• Pandas, NumPy

📚 Dataset Sources

• ChestX-ray14 (NIH)
• COVID-19 Radiography Dataset
• Pneumonia Dataset

🙋‍♂️ Author

Saravanan
📧 [email protected]
🌐 LinkedIn Profile

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.