FRANÇAIS
A showcase of ÉTS researchers’ publications and other contributions
SEARCH

RDO cost modeling for low-complexity HEVC intra coding

Downloads

Downloads per month over past year

Jamali, Mohammadreza et Coulombe, Stéphane. 2016. « RDO cost modeling for low-complexity HEVC intra coding ». In IEEE Canadian Conference on Electrical and Computer Engineering (CCECE) (Vancouver, BC, Canada, May 15-18, 2016) Piscataway, NJ, USA : IEEE.
Compte des citations dans Scopus : 1.

[img]
Preview
PDF
Coulombe S. 2016 14174 RDO Cost Modeling for Low-Complexity.pdf - Accepted Version
Use licence: All rights reserved to copyright holder.

Download (527kB) | Preview
[img]
Preview
PDF
Coulombe S. 2016 14174 poster RDO Cost Modeling for Low-Complexity.pdf - Presentation
Use licence: All rights reserved to copyright holder.

Download (954kB) | Preview

Abstract

High efficiency video coding (HEVC) is the newest international standard for video compression, providing improved coding performance that achieves compression ratios up to 50% higher than those obtained with H.264/AVC. However, this improvement comes at the expense of high computational complexity and coding time. In this paper, we propose a novel method for fast and low-complexity intra HEVC mode decision based on rate-distortion optimization (RDO) cost modeling, which permits the exclusion of non-promising candidates from the RDO processing. To achieve even more complexity reduction, an additional rough most probable modes examination is coupled with the main algorithm. Experimental results show that the proposed algorithms reduce the encoding time by 41.8% on average, with a negligible quality loss of 0.058 dB (BD-PSNR) for all-intra scenarios, as compared to the HEVC reference implementation, the HM 15.0.

Item Type: Conference proceeding
Professor:
Professor
Coulombe, Stéphane
Affiliation: Génie logiciel et des technologies de l'information
Date Deposited: 16 Dec 2016 15:45
Last Modified: 30 May 2017 18:52
URI: https://espace2.etsmtl.ca/id/eprint/14174

Actions (login required)

View Item View Item