Efficient Representation for Point Clouds: Compression, Processing and Distortion Quatification

Written by  |  23/08/2021 - 08/31


Introduction:

Having its outstanding flexibility for representing 3D objects realistically and naturally, Point Cloud (PC) has become a popular media format used in vast applications, such as the Augmented/Virtual Reality (AR/VR), autonomous driving, cultural e-heritage, and digital twin, for immersive service enabling. It then urgently calls for high-efficiency compression of point cloud, both lossy and losslessly.


A point cloud is a collection of non-uniformly and sparsely distributed points that are characterized using their 3D coordinates (e.g., (x,y,z)) and  attributes (e.g., RGB colors, reflectances) if applicable. Unlike those well-structured pixel grids of a 2D image plane or video frame, point cloud relies on unconstrained displacement of points to freely represent arbitrary-shaped 3D objects. However, this puts obstacles on efficient coding of geometric occupancy due to the difficulty in characterizing and exploiting inter-correlation across spatially scattered points in a 3D space.


Talks:

ICIG 2021


Publication:

  • Wenjie Zhu, Yiling Xu, Dandan Ding, Mike Nilsson, and Zhan Ma, "Lossy Point Cloud Geometry Compression via Region-wise Processing," accepted by IEEE Trans. Circuits and Systems for Video Technology, July 2021.

  • Linyao Gao, Tingyu Fan, Jianqiang Wang, Yiling Xu, and Zhan Ma, "Point Cloud Geometry Compression via Neural Graph Sampling," accepted by IEEE ICIP, May 2021.

  • Wenjie Zhu, Zhan Ma, Yiling Xu, Li Li, and Zhu Li, "View-Dependent Dynamic Point Cloud Compression,"  IEEE Trans. Circuits and Systems for Video Technology, vol. 31, no. 2, pp. 765-781, Feb. 2021.

  • Jianqiang Wang, Hao Zhu, Haojie Liu and Zhan Ma, "Lossy Point Cloud Geometry Compression via End-to-End Learning," accepted by IEEE Trans. Circuits and Systems for Video Technology, Jan. 2021. 

  • Jianqiang Wang, Dandan Ding, Zhu Li, and Zhan Ma, "Multiscale Point Cloud Geometry Compression," IEEE DCC, Snow Bird, Utah, March 2021

  • Qi Yang, Zhan Ma, Yiling Xu, Zhu Li and Jun Sun, "Inferring Point Cloud Quality via Graph Similarity," accepted by IEEE Trans. on Pattern Analysis and Machine Intelligence, Dec. 2020. 

  • Qi Yang, Hao Chen, Zhan Ma, Yiling Xu, Rongjun Tang, and Jun Sun, "Predicting the Perceptual Quality of Point Cloud: A 3D-to-2D Projection-Based Exploration," accepted by IEEE Trans. Multimedia, Sept. 2020