NeuralViewSynthesisandMatching forSemi-SupervisedFew-ShotLearningof3DPose

Neural Information Processing Systems 

Ourmodel is trained in an EM-type manner alternating between increasing the 3D pose invariance ofthefeature extractor andannotating unlabelled data through neural viewsynthesis andmatching.

Similar Docs  Excel Report  more

TitleSimilaritySource
None found