I have trained a DCNN on lung images using image labeler obtained data. But instead of getting the whole image with lungs and background, I would like to segment out the lungs region and discard the background. Any idea on how to do it?
You cannot use the Image Labeler to segment/crop out the image using an ROI. You can, however, simply use the imcrop function to crop the images using their bounding boxes declared in the Image Labeler app.