Multi-Modal-Alignment

kVCT–MVCT Alignment 3D-Slicer

This repository contains research scripts for automating two tasks inside 3D Slicer:

1. Alignment → Align kVCT (moving) to MVCT (fixed) volumes
2. Body Mask Extraction → Extract binary BODY mask from MVCT segmentations

⚠️ Disclaimer: These scripts are for research purposes only. Not for clinical use.

---

📂 Repository Structure

kvct-mvct-alignment/
├─ scripts/
│  ├─ slicer_alignment.py       # alignment automation script
│  └─ slicer_bodymask.py        # body mask extraction script
├─ docs/
│  ├─ usage_alignment.md        # detailed alignment steps
│  ├─ usage_bodymask.md         # detailed mask extraction steps
│  └─ overview.md               # optional workflow notes
├─ examples/
│  ├─ example_scene.png
│  └─ example_output_mask.npy
├─ .gitignore
├─ LICENSE
└─ README.md

---

🚀 Quick Start

1. Open 3D Slicer
2. Load patient scene (kVCT + MVCT)
3. Open the Python Interactor (Ctrl+3)
4. Run one of the provided scripts:
   • scripts/slicer_alignment.py → align kVCT to MVCT
   • scripts/slicer_bodymask.py → extract BODY mask
5. See detailed instructions in the docs/ folder

---

🔄 Workflow Overview

The following diagram shows the full alignment process from kVCT to MVCT:

🩻 Workflow 1 — Alignment (kVCT → MVCT)

This script automates alignment of kVCT to MVCT inside 3D Slicer.

Steps

1. Load Patient Data

   • Open patient scene in Slicer
   • Verify volumes:
     • kct_p1 (moving)
     • mct_p1 (fixed)

2. Update Script Variables

   patientID = '01'            # Patient no:1 - Folder Name
   kctVolumeName = 'kct_p1'    # Patient no:1 kVCT modality
   mctVolumeName = 'mct_p1'    # Patient no:1 MVCT modality
   path2SaveAligned = r'C:\Folder_Path'

3. Resample kVCT

   • Module: Resample Scalar Volume
   • Spacing: (0.754, 0.754, 2.0)
   • Interpolation: BSpline
   • Save as: kct_p1_resliced

4. Crop Volume

   • Module: Crop Volume → “Fit to Volume”
   • Save as: kct_p1_resliced_cropped

5. Apply Transform

   • Module: Transforms → add Linear Transform
   • Assign to kct_p1_resliced_cropped
   • Adjust → Harden Transform

6. Save Outputs

   • kctAligned_<patientID>.npy → aligned kVCT
   • final_transform.tfm → transform parameters
   • Scene file (File → Save, check all)

---

🩺 Workflow 2 — Body Mask Extraction (MVCT → Mask)

This script extracts a binary body mask from MVCT segmentations inside 3D Slicer.

Steps

1. Load Patient Scene

   • Open from CRO_Dataset
   • Verify MVCT segmentation exists

2. Update Script Variables

   patientID = '58'
   segmentationName = '814102716: RTSTRUCT: CT_1'  # segmentation node in Slicer
   segmentName = 'BODY'                           # segment to extract
   path2SaveMask = r'C:\Folder_Path'

3. Prepare Output Folder

   • Ensure kctAlignedMask folder exists at:
     C:\Mubashara\Dataset_Working\AI_Dataset_UNIUD

4. Run Script

   • Open Python Interactor (Ctrl+3)
   • Paste + run script

5. Verify Outputs

   • BODY mask saved as .npy in path2SaveMask
   • Scene updated with segmentation nodes

6. Save Scene

   • File → Save → check all boxes → confirm

---

📊 Outputs

• Alignment

  • kctAligned_<patientID>.npy → aligned kVCT
  • final_transform.tfm → transform parameters
  • Updated Slicer scene

• Body Mask Extraction

  • <patientID>_BODY_mask.npy → binary body mask

---

🖼 Example

Alignment result in 3D Slicer slice views:

---

🚫 Data Privacy

• Do not commit real patient DICOM/NIfTI volumes
• .gitignore already excludes sensitive formats:

  *.nii
  *.nii.gz
  *.nrrd
  *.mha
  *.mhd
  *.dcm
  *.npy
  outputs/
  

---