Paper Info
Title
RocNet: Recursive Octree Network for Efficient 3D Deep Representation
Abstract
We introduce a deep recursive octree network for the compression of 3D voxel data. Our network compresses a voxel grid of any size down to a very small latent space in an autoencoder-like network. We show results for compressing 32 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> , 64 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> and 128 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> grids down to just 80 floats in the latent space. We demonstrate the effectiveness and efficiency of our proposed method on several publicly available datasets with three experiments: 3D shape classification, 3D shape reconstruction, and shape generation. Experimental results show that our algorithm maintains accuracy while consuming less memory with shorter training times compared to existing methods, especially in 3D reconstruction tasks.
Year
DOI
Venue
2020
10.1109/3DV50981.2020.00051
2020 International Conference on 3D Vision (3DV)
Keywords
DocType
ISSN
RocNet,efficient 3D deep representation,deep recursive octree network,3D voxel data,voxel grid,3D shape classification,3D shape reconstruction,shape generation,autoencoder-like network
Conference
2378-3826
ISBN
Citations 
PageRank 
978-1-7281-8129-5
0
0.34
References 
Authors
0
3
Name
Order
Citations
PageRank
Juncheng Liu1146.66
Steven Mills24117.74
brendan mccane322333.05