Thanks for joining us for SIGGRAPH 2020. Will we see you at SIGGRAPH 2021?

Technical Paper
:
CNNs on Surfaces Using Rotation-equivariant Features
Event Type
Technical Paper
Registration Levels
US
UA
EU
Time
Location
DescriptionWe propose an approach for CNNs on surfaces, addressing the fundamental problem of rotational ambiguity in geometric deep learning. The solution combines rotation-equivariant learning and vector-valued features.