Output of segmentation network

[pvnieo]

Hi,

Thanks for the great contribution and for releasing the code.

I'm not having an issue, but a question, before investing time into integrating this network into my code.

For the segmentation case, does the output features are per vertex, or per face of the input shape?

Thank you in advance!

[pvnieo]

Hi, In the segmentation scenario, the network outputs a feature per edge to predict the class of the edge. Is there a way to aggregate these feature to produce a point based features?

Thank you in advance!