3d visualizations of MRI data

2 min read 2025-03-31

1 Introduction

I found 3d visualization is really useful for understanding MRI results. Here is a brief guide which is mainly based on this paper.

2 Convert nii to VTK

First of all, we should convert nii files to VTK format using ITK-SNAP.

  1. Open a main image (e.g. whole brain mask).
  2. Open a segmentation (e.g. ROIs).
  3. Export as surface mesh and select labels.
  4. Save as “VTK PolyData File”. Separate files will be named as “prefix+value”.
  5. A 3d view will be open in the bottom-left window after saving.

3 Render in 3d

ParaView is an open source visualization application which could render VTK files as surface in 3d.

  1. Open VTK files, and press Apply (Open as a group will generate time series).
  2. Show all in the pipeline browser.
  3. Select every file and add a Group Datasets filter.
  4. Change properties.
  5. Save state.

4 Filtering data

In ParaView, filters are pipeline modules or algorithms that have inputs and outputs. Here are some useful filters.

  • GroupDatasets: groups multiple datasets to create a single data set (without actually duplicate the heavy data)
  • Glyph: place markers or glyphs at point locations in the input dataset.
  • Threshold: cells that have scalars within the specified range (include the lower and upper threshold) will be passed through the filter.
  • Clip: clip any dataset using either an implicit function (such as a plane, sphere, or a box)
  • Calculator: computes a new data array or new point coordinates as a function of existing input arrays. The name of variable is case sensitive. Results can be seen in the spreadsheet view.