Goto

Collaborating Authors

 Asia



Augmentation-Aware Self-Supervision for Data-Efficient GAN Training

Neural Information Processing Systems

We further encourage the generator to adversarially learn from the self-supervised discriminator by generating augmentation-predictable real and not fake data.








PhysGNN: APhysics-DrivenGraphNeuralNetwork BasedModelforPredictingSoftTissueDeformationin Image-GuidedNeurosurgery

Neural Information Processing Systems

On the other hand, model-based methods are a set of registration techniques that treat images as a deformable volume--a notion first introduced by Broit [1981]--to better allow for presenting elastic and plastic deformations.