Professional MPR Viewer with AI

A modern Medical Image Viewer built with PyQt5, NumPy, Nibabel, and TensorFlow that supports visualization of NIfTI and DICOM volumes in Axial, Sagittal, Coronal, and Oblique planes.
It also includes an AI-based orientation classifier and tools for measurement, ROI extraction, and mask overlay.

---

🧠 Features

• Multi-Planar Reconstruction (MPR): Axial, Sagittal, Coronal, and Oblique views.
• AI Classification: Automatically classifies NIfTI file orientation (Axial, Coronal, Sagittal).
• DICOM to NIfTI conversion (via dicom2nifti).
• Brightness and Contrast control per view.
• Zoom and playback controls (cine mode).
• ROI selection and export as new NIfTI files.
• Measurement tool for pixel distance.
• Mask overlay and contour visualization.
• Resizable interface with a modern dark QSS theme.

---

⚙️ Requirements

Install the required dependencies using pip:

pip install PyQt5 numpy nibabel dicom2nifti pydicom scikit-image scipy tensorflow matplotlib

⚠️ If you face missing-library errors, the app will display a message with the correct install command.

---

📁 File Structure

main.py              # Main GUI and logic
nifti_orientation_classifier_full_model.h5   # (Optional) Default AI model file

---

🚀 Usage

1. Run the app:

   python main.py

2. Load Data:

   • NIfTI File: Click “Load NIfTI” and select a .nii or .nii.gz file.
   • DICOM Folder: Click “Load DICOM (.dcm)” and choose a folder containing .dcm files.
     They will be converted automatically to NIfTI.

3. AI File Classifier:

   • Click “Classify NIfTI File” to detect its anatomical orientation.
   • You can switch models using “Change AI Model.”

4. Manipulate Views:

   • Adjust tilt angles in Oblique Plane Controls.
   • Zoom, pan, or modify brightness and contrast interactively.
   • Enable Measurement Mode to measure distances.

5. ROI Tools:

   • Select slice ranges and apply or export ROI as a new NIfTI file.
   • Reset to restore the full volume.

6. Mask Overlay:

   • Load mask .nii files and visualize outlines or surface boundaries.

---

🧩 Notes

• The AI model path is defined in:

  self.model_path = r"D:\Programing\Biomedical Projects\Task 2\Files\Ai Model Weight\nifti_orientation_classifier_full_model.h5"

  Update it to your local model file path if needed.

• Temporary converted or exported files are saved under:

  D:\Programing\Biomedical Projects\Task 2\Files\Temp
  

• If you open large datasets, initial loading may take a few seconds.

---

🧰 Tech Stack

• Python 3.8+
• PyQt5 — GUI and visualization
• NumPy / SciPy — numerical operations
• Nibabel — NIfTI handling
• dicom2nifti / pydicom — DICOM conversion
• scikit-image — contour extraction
• TensorFlow / Keras — AI model inference

---

🩺 Screenshot

---

📜 License

This project is for educational and research purposes only.
Not intended for clinical use.