liver segmentation


(Marco Festugato) #1

Hi guys, I have to segment the liver in 199 different images of the same patient, any ideas/code/example about the classes/methods i should use?
Thanks in advance!


(Dženan Zukić) #2

In 199 slices of the same 3D volume, or 199 different 3D volumes? If it is the first, I recommend that you use ITK-SNAP or 3D Slicer and do the segmentation semi-automatically. If it is the second, you should look into algorithms developed for that purpose, e.g. here. You can find more liver-related challenges here.


(Marco Festugato) #3

Hi dzenanz and thank u for your kind answer. I have 199 DICOM images (of CT abdomen) from the same patient.

I first want to roughly segment the liver because what I really wanna do is the segmentation of the liver vessels… I was thinking about using the confidence connected to segment the liver, what do u think?
Unfortunately for my project ITK is mandatory, so i cant use 3D slicer…
Thanks in advance!


(Dženan Zukić) #4

That sounds like a school assignment! How about you try a few algorithms and see which one works best?

Look at this example to see how to read a DICOM series. And this example demonstrates how to directly apply a segmentation algorithm. I guess you will have to look at other examples to find one which can accomplish your task:


(Marco Festugato) #5

Hi dzenanz and thank u for your answer! I’ll definitely do it :slight_smile: