Professional MPR Viewer with AI Orientation Prediction

A medical imaging application for Multi-Planar Reconstruction (MPR) with AI-powered orientation detection.

Features

• 4 Synchronized Views: Axial, Sagittal, Coronal, and Oblique planes
• Interactive Navigation: Drag crosshairs to navigate through 3D volumes
• AI Orientation Prediction: Automatically detect view orientation
• Organ Detection: Identify anatomical regions from DICOM metadata
• Surface Boundary Mode: Display 2D contours from segmentation masks
• ROI Export: Save regions of interest as NIfTI files

Quick Start

Installation

# Clone this repository
git clone https://github.com/Marize-Raafat/Medical-image-MPR-viewer.git
cd mpr-viewer

# Install dependencies
pip install -r requirements.txt

# Run the application
python src/MPR_viewer_final.py

Usage

1. Click "Load NIfTI" or "Load DICOM" to open a medical image
2. Drag the colored crosshairs to navigate
3. Use Oblique Tilt sliders to adjust the oblique plane
4. Click "Predict Orientation" to run AI classification
5. Click "Detect Organ" (DICOM only) to identify anatomy

AI Model

The orientation prediction model requires:

• models/config.json - Model architecture
• models/model.weights.h5 - Trained weights

Note: Model files not included due to size.

Requirements

• Python 3.8+
• 8GB RAM minimum
• GPU optional (CPU inference supported)

Project Structure

mpr-viewer/
├── src/                    # Source code
├── models/                 # AI model files
├── docs/                   # Documentation
└── requirements.txt        # Python dependencies

License

This project is licensed under the MIT License.

Author

Marize Raafat

• GitHub: @Marize-Raafat
• Email: marizeraafat020@gmail.com