[Feature Request] Add high level segmentation (not semantic)

[thhart]

BoofCV has a lot of low level segmentation tools which are good already. However I am missing a more high level approach towards semantic segmentation. This could be used to process further with NN leading to semantic approaches. Do you think this would be in the scope of BoofCV yet? IMHO BoofCV could build a bridge here. But this is only my very personal view.

Maybe worth to read this whoever is interested:

https://stackoverflow.com/questions/33947823/what-is-semantic-segmentation-compared-to-segmentation-and-scene-labeling

[lessthanoptimal]

Like superpixels? That's in BoofCV and they were all the rage before deep learning took over.