Image Segmentation question

I am wondering if anyone has had any luck on doing image segmentation problems on their own datasets (i.e. not the camvid dataset). I am wondering if whoever has had luck, how did you go about it? Especially with regards to annotating your own images. It’s not really discussed too much in part 1 of the course. The purpose of this question–yes, I know I can just Google how to do it, but not everyone knows about Fastai and I really would like some advice on how to “go deep” into this topic from people who are from the Fastai ecosystem