Visualization of connected parts of a structure

Hello guys:

I would like to ask a question about the visualization of a 3D structure. I want to use different colors to show the connected parts of a structure reconstructed from micro-CT images. Is there any good ways to achieve that?


Here is my link for the
https://drive.google.com/drive/folders/1PSyQPIZa7UFECnDjPQdvgliUmlayApm2?usp=sharing

Hi,

MorphoLibJ is the great tool kit for this type of analysis (can do 3D too). Install it from the Help > Update... sites and select IJPB plugins.

What you want is a connected component labelling, Plugins > MorphoLibJ > Binary Images > Connected Components Labelling. This fills each connected regions with an integer starting with 1 and goes up, which makes it hard to see.

So apply a color labelling using Plugins > MorphoLibJ > Label Images > Set Label Map. Choose Golden Angle, try the shuffle and set Black Background

There bunch more things in MorphoLibJ that you might find useful. For example:

  • removing small areas with size Opening
  • closing small gaps with a morphological closing
  • measuring the geodesic distance between regions.

Cheers,

Chris

2 Likes

Hi Chris:

Thanks a lot! I did what you stated on 20 images but the connected components in the final map are so sparse. The connectivity values of 6 and 26 have been tried in my case. Is there any trick I don’t know?

Best,
Zhazha

Hi,

Check that the values inside the mask areas are 255 and the background is 0. When I opened your image I had to invert then and invert the the LUT. Sometimes Fiji gets mixed up when importing TIFFs.

Cheers,

Chris

Hi Chris:

Yes. It is this problem. Do you know the value of connectivity? I think you set with a higher value, right? What is the meaning of the value?

Best,
Zhazha

Hi,

In 2D:

  • 4-connected means regions must touch by at least 1 pixel edge
  • 8-connect means touching by at least 1 corner.

In 3D:

  • 6-connect is one voxel edge
  • 26-connected is one voxel corner.

Higher connectivity means less regions.

Cheers,

Chris

Thanks very much! I get a new skill!