I thought that too. Just finished a kaggle competition involving segmentation, like a lot of participant I used one form of U-net (my own implementation).

You can probably find a lot of u-net implementations from this contest. One that performed really well [1]. It uses 'inception style' blocks feature extraction instead of vgg. But otherwise pretty similar.

[1] https://github.com/EdwardTyantov/ultrasound-nerve-segmentati...