International Display Workshops General Incorporated Association

14:00 〜 14:20

[VHF1/DES1-4] Deep Learning Based Layered Point Cloud Compression for Representing Shape Aware Level of Detail

*Hideaki Kimata1 (1.Kogakuin University (Japan))

point cloud compression, deep learning, level of detail

https://doi.org/10.36463/idw.2022.0645

Point cloud compression methods based on deep learning are being studied. Extracting a part of point cloud from compressed data to grasp the characteristic shape of an object is an important function. I propose ranking and hierarchical coding that can sequentially decode main features of point cloud.