ITK: best way to apply a filter to part of an image

Question

I am an ITK newbie, so please excuse any misused jargon. I have a series of JPEG files which show screenshots from ultrasound imaging software. I'd like to find just the ultrasound part of the screen (specifically NOT a square), then run that part of the image through some filters, leaving the rest of the screen untouched. I imagine I can use a binary mask filter to extract the area of interest, process the masked image, and then use a combination of mask and addition to get the two parts back together. Is there a better way to do this?

Answer 1

You may create a binary mask by hand of your region of interest. A very simple tool for this is ITK-Snap.

If you are not interested in certain pixels of a certain intensity / color, you may use a pipeline of BinaryThreshold filters

Answer 2

These itkUltrasoundProbeGeometryCalculator classes where written to find the sector portions of an ultrasound image screenshot for extraction and further analysis:

https://github.com/KitwareMedical/TubeTK/blob/master/Base/USTK/itkUltrasoundProbeGeometryCalculator.h