This might be more of an issue for scikit-image, but it would be easy to add here: if the skeletonize widget receives an integer-valued input, it currently binarizes it and returns a {0, 1} valued skeleton. Instead, we could multiply the input image with the binary skeleton, so that each skeleton would have the same value as the original labels image. This might be less surprising behaviour to users because the labels and the skeletonized labels would have the same colour scheme.
This might be more of an issue for scikit-image, but it would be easy to add here: if the skeletonize widget receives an integer-valued input, it currently binarizes it and returns a {0, 1} valued skeleton. Instead, we could multiply the input image with the binary skeleton, so that each skeleton would have the same value as the original labels image. This might be less surprising behaviour to users because the labels and the skeletonized labels would have the same colour scheme.
CC @jamesyan-git