Unsupervised Point Cloud Pre-training via Occlusion Completion.

Hanchen Wang Qi Liu Xiangyu Yue Joan Lasenby Matt J. Kusner

Published in: ICCV (2021)

Keyphrases

point cloud
structure from motion
supervised learning
surface reconstruction
point sets
missing parts
stereo camera
laser scanner
training set
point cloud data
semi supervised
triangular mesh
feature vectors
implicit surfaces
high quality
three dimensional
machine learning