Fast Texture Intra Size Coding Based On Big Data Clustering for 3D-Hevc
Hamza Hamout, Abderrahmane Elyousfi
Abstract
High Efficiency Video Coding (HEVC) based 3D video coding (3D-HEVC) is the latest new joint standardization effort of ISO/IEC MPEG and ITU - T Video Coding Experts Group (VCEG) for 3D video coding. This new standard provides a significant coding improvement, especially for high resolution videos. However, one the most important challenges in 3D-HEVC is time complexity. In technical terms, 3D-HEVC is a hybrid video coding approach using quad-tree based block partitioning with more flexible Coding Unit (CU) size selection, which increases the coding efficiency of 3D-HEVC significantly, but also brings huge computational complexity due to the Rate Distortion (RD) cost calculation of all possible dimensions of CU to select the optimal one. To reduce this computational complexity, this paper proposes an efficient fast intra coding unit based on big data analysis. The experimental results demonstrate that the novel approach provides a significant trade-offs between computational complexity and RD performance.
BibTeX
@inproceedings{icassp2018_fasttextureintra,
title = {Fast Texture Intra Size Coding Based On Big Data Clustering for 3D-Hevc},
author = {Hamza Hamout and Abderrahmane Elyousfi},
booktitle = {ICASSP 2018},
year = {2018}
}