We propose a Point-Voxel DeConvolution (PVDeConv) module for 3D dataautoencoder . To demonstrate its efficiency we learn to synthesize high-resolution point clouds of 10k points that densely describe the underlyinggeometry of Computer Aided Design (CAD) models . We introduce a dedicated dataset, the CC3D, containing 50k+ pairs of CAD models and their corresponding 3D meshes . This dataset is used to learn a convolutionalautoincoder for point clouds sampled from the pairs of 3D scans – CAD models . The challenges of this new dataset are demonstrated in comparison with othergenerative point cloud sampling models trained on ShapeNet . We are efficient with respect to memory consumption and training time .

Author(s) : Kseniya Cherenkova, Djamila Aouada, Gleb Gusev

Links : PDF - Abstract

Code :
Coursera

Keywords : point - d - models - cad - dataset -

Leave a Reply

Your email address will not be published. Required fields are marked *