Open af3113 opened 6 years ago
@af3113 Actually the two digit images segmented as showed in this README file are got by reconstructing from two longest DigitCap vectors. I think if the CapsNet really works as what Hinton wants, the coupling coefficient will imply the segmentation.
Hello,
Thanks for sharing your code - very helpful indeed to start using CapsNet. I wonder how you got your segmentations? I am asking as I would like to use CapsNets for semantic segmentation. Any ideas?
Many thanks!