I am new to the Medical Image Processing area. I have a project about lung segmentation using ITK and C++. I am using a set of chest CT scans on the LOLA11 dataset, you can easily download these CT images from their website. Basically, I need to apply the best lung segmentation method to these dataset in order to find the volume of each lung. I will be more than happy if you help me find the code for the segmentation. Thank you.
Asked
Active
Viewed 571 times
-1
-
2Please read this: [Stack Overflow: How do I ask a good question?](http://stackoverflow.com/help/how-to-ask) – paulsm4 Apr 05 '16 at 02:48
-
Give a link for the images so we can see them easily. – Malcolm McLean Oct 06 '16 at 12:41
1 Answers
1
This is a redacted copy of Luis Ibanez's answer on ITK mailing list:
- Extract a region of interest that includes the trachea and the
full lungs, but not the head of the patient (e.g. cut below the
head, somewhere in the neck). You can use the
RegionOfInterestImageFilterfor this. - Use the
ConnectedThresholdImageFilter - Set seed points inside the trachea (one seed point should be enough)
- Set the lower threshold to -700 Hounsfield Units
- Set the upper threshold to -200 Hounsfield Units
- Run the filter. This should give you a binary mask of the two lungs. If you find many holes inside the mask, you could clean it up with the Morphological filters or the Voting Hole filling filters.
Paul Jurczak
- 7,008
- 3
- 47
- 72