Point cloud compression framework for the web

This work is dedicated to the compression of 3D point clouds in order to allow an efficient and quick transmission of point cloud datasets (PCD) for visualization over the internet. Standard methods include quantization of vectors or cloud simplification via octree structures. While a quantization i...

Full description

Saved in:
Bibliographic Details
Published in:International Conference on 3D Imaging pp. 1 - 8
Main Authors: Renault, Sylvain, Ebner, Thomas, Feldmann, Ingo, Schreer, Oliver
Format: Conference Proceeding
Language:English
Published: IEEE 01.12.2016
Subjects:
ISSN:2379-1780
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This work is dedicated to the compression of 3D point clouds in order to allow an efficient and quick transmission of point cloud datasets (PCD) for visualization over the internet. Standard methods include quantization of vectors or cloud simplification via octree structures. While a quantization into a bit representation will transform vectors in discrete positions, we added octree structures for a fixed level for re-indexing the quantized points relative to the individual local position of the subdivisions. Each subdivision multiplies the resolution of one coordinate by two, by adding a "virtual" bit. This bit can be then removed from the quantization bytes. So the combination of quantization and fixed octree structures decreases the amount of needed quantization bits without losing resolution. On the contrary, it is possible to increase the resolution of a PCD by adding the "virtual" bit to the quantized data without significantly changing the size of the dataset. We demonstrated the feasibility of this technique for the web by developing a lightweight framework, running in a browser-based environment.
ISSN:2379-1780
DOI:10.1109/IC3D.2016.7823455