MediScan: AI-Powered Bone Fracture Detection

</div> <div align="center">

</div> <p align="center"> <i>Advanced deep learning system for rapid and accurate bone fracture detection from X-ray images.</i> </p>

📋 Overview

MediScan is a state-of-the-art web application that uses artificial intelligence to help healthcare professionals identify bone fractures from X-ray images with high accuracy. The system leverages YOLOv8 deep learning models and provides visual explanations using Grad-CAM technology to enhance trust and interpretability.

✨ Key Features

🔍 Advanced Detection: Identifies various fracture types (transverse, oblique, spiral, comminuted, greenstick)
⚡ Speed: Processes X-ray images in under 2 seconds
🔬 Grad-CAM Visualization: Shows exactly where the model is focusing, increasing clinical trust
📊 Detailed Analysis: Provides confidence scores and explanations for each detection
🌐 Web Interface: User-friendly frontend for easy image uploads and result viewing
🔄 API Access: RESTful endpoints for integration with other clinical systems

🖼️ Screenshots

<div align="center"> <table> <tr> <td><strong>Detection Results</strong></td> <td><strong>Grad-CAM Visualization</strong></td> </tr> <tr> <td><img src="frontend/images/sample-result.jpg" alt="Detection Results" width="100%"/></td> <td><img src="frontend/images/sample-gradcam.jpg" alt="Grad-CAM Visualization" width="100%"/></td> </tr> </table> </div>

🏗️ Architecture

MediScan consists of two main components:

Backend (FastAPI): Handles image processing, runs the YOLOv8 model, and generates visualizations
Frontend (HTML/CSS/JS): Provides user interface for uploading X-rays and viewing results

🛠️ Technology Stack

Backend: Python, FastAPI, PyTorch, Ultralytics YOLOv8, OpenCV
Frontend: HTML5, CSS3, JavaScript, FontAwesome
AI Model: YOLOv8 trained on bone fracture X-ray datasets
Visualization: Grad-CAM (Gradient-weighted Class Activation Mapping)

📂 Project Structure

├── graphs/
│   ├── epochs_vs_accuracy.png
│   └── confusion_matrices/
├── models/
│   ├── yolo_model.py
├── backend/                # FastAPI backend
│   ├── app.py              # Main application file
│   ├── requirements.txt    # Backend dependencies
│   ├── models/             # YOLOv8 model files
│   ├── results/            # Detection results
│   │   ├── explanations/   # Model explanation images
│   │   └── gradcam/        # Grad-CAM visualization images
│   └── uploads/            # Uploaded X-ray images
│
├── frontend/               # Static web frontend
│   ├── index.html          # Home page
│   ├── detect.html         # Detection page
│   ├── gradcam.html        # Grad-CAM explanation page
│   ├── about.html          # About page
│   ├── features.html       # Features page
│   ├── team.html           # Team page
│   ├── contact.html        # Contact page
│   ├── styles.css          # Main stylesheet
│   └── images/             # Frontend images
│
├── requirements.txt        # Project dependencies
└── README.md               # Project documentation

🚀 Getting Started

Prerequisites

Python 3.8+
pip (Python package manager)
Modern web browser

Installation

Clone the repository

git clone https://github.com/yourusername/mediscan.git
cd mediscan

Install dependencies
```
pip install -r requirements.txt
```
Start the backend server
```
cd backend
uvicorn app:app --reload
```
Open the frontend
- Navigate to the frontend folder
- Open index.html in your web browser
OR
- Serve the frontend using a simple HTTP server:
```
cd frontend
python -m http.server 8080
```
- Open http://localhost:8080 in your browser

📡 API Endpoints

| Endpoint | Method | Description | |----------|--------|-------------| | / | GET | API status check | | /status | GET | Get backend system status | | /detect | POST | Upload and analyze X-ray image | | /gradcam/{image_id} | GET | Get Grad-CAM visualization for a specific image |

Example API Usage

import requests

# Upload an X-ray image for detection
with open('xray.jpg', 'rb') as f:
    files = {'file': f}
    response = requests.post('http://localhost:8000/detect', files=files)

result = response.json()
print(f"Detection ID: {result['detection_id']}")
print(f"Result image: {result['result_image']}")
print(f"Grad-CAM visualization: {result['gradcam_image']}")

💻 Development Setup

For developers who want to contribute to the project:

Create a virtual environment

python -m venv env
source env/bin/activate  # On Windows: env\Scripts\activate

Install development dependencies
```
pip install -r requirements.txt
```

Run backend with debug mode

cd backend
uvicorn app:app --reload --debug

👥 Our Team

MediScan is developed by a multidisciplinary team of AI/ML developers and software engineers:

Alok Kumar - Full Stack Dev - Lead developer specializing in AI integration and full-stack development
Amrit Kumar - Technical Lead - Project architect and technical lead overseeing system design and implementation
Aashish Kumar - Python Dev - Backend Python developer specializing in AI model integration and machine learning
Samridhi Bagchi - Frontend Developer - UI/UX specialist focused on creating intuitive medical interfaces
Karan Singh - ML Engineer - AI/ML expert specializing in medical image processing and deep learning model optimization

📜 License

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

🙏 Acknowledgements

Ultralytics YOLOv8 for the object detection model
FastAPI for the backend framework
Medical partners for providing expertise and testing

MediScan

Install / Use

README