A fast hybrid decision algorithm for H.264/AVC intra prediction based on entropy theory

Guifen Tian, Tianruo Zhang, Takeshi Ikenaga, Satoshi Goto

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Rate Distortion Optimization based spatial intra coding is a new feature of H.264/AVC standard. It efficiently improves the video coding performance by brutally utilizing variable block sizes and multiple prediction modes. Thus, extremely high computation complexity is required. This paper proposes a hybrid decision algorithm which can reduce unimportant block sizes and prediction modes. Entropy feature of each MB is extracted to decide the optimal block size. Then Prewitt operators based edge direction extraction module is designed to select only promising prediction modes. In addition, above two decision modules can function together to form a fast hybrid decision algorithm for intra prediction. Sufficient experiments prove that this hybrid algorithm is able to achieve 63% time reduction with PSNR loss limited to 0.07dB, bit rate increase at most 2.02%. It is expected to be utilized as a favorable accelerator for a real-time H.264 encoder for HDTV.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages85-95
Number of pages11
Volume5371 LNCS
DOIs
Publication statusPublished - 2009
Event15th International Multimedia Modeling Conference, MMM 2009 - Sophia-Antipolis
Duration: 2009 Jan 72009 Jan 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5371 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other15th International Multimedia Modeling Conference, MMM 2009
CitySophia-Antipolis
Period09/1/709/1/9

Fingerprint

Intra Prediction
Entropy
Prediction
Rate-distortion Optimization
Module
Video Coding
Hybrid Algorithm
Encoder
Accelerator
High definition television
Coding
Image coding
Sufficient
Particle accelerators
Real-time
Operator
Experiment
Experiments

Keywords

  • Block Size Decision
  • Entropy Theory
  • H.264/AVC
  • Intra Prediction
  • Mode Decision

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Tian, G., Zhang, T., Ikenaga, T., & Goto, S. (2009). A fast hybrid decision algorithm for H.264/AVC intra prediction based on entropy theory. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5371 LNCS, pp. 85-95). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5371 LNCS). https://doi.org/10.1007/978-3-540-92892-8_10

A fast hybrid decision algorithm for H.264/AVC intra prediction based on entropy theory. / Tian, Guifen; Zhang, Tianruo; Ikenaga, Takeshi; Goto, Satoshi.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5371 LNCS 2009. p. 85-95 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5371 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Tian, G, Zhang, T, Ikenaga, T & Goto, S 2009, A fast hybrid decision algorithm for H.264/AVC intra prediction based on entropy theory. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 5371 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5371 LNCS, pp. 85-95, 15th International Multimedia Modeling Conference, MMM 2009, Sophia-Antipolis, 09/1/7. https://doi.org/10.1007/978-3-540-92892-8_10
Tian G, Zhang T, Ikenaga T, Goto S. A fast hybrid decision algorithm for H.264/AVC intra prediction based on entropy theory. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5371 LNCS. 2009. p. 85-95. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-540-92892-8_10
Tian, Guifen ; Zhang, Tianruo ; Ikenaga, Takeshi ; Goto, Satoshi. / A fast hybrid decision algorithm for H.264/AVC intra prediction based on entropy theory. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5371 LNCS 2009. pp. 85-95 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{df3d8aae253f47af98e5cbe0ee736524,
title = "A fast hybrid decision algorithm for H.264/AVC intra prediction based on entropy theory",
abstract = "Rate Distortion Optimization based spatial intra coding is a new feature of H.264/AVC standard. It efficiently improves the video coding performance by brutally utilizing variable block sizes and multiple prediction modes. Thus, extremely high computation complexity is required. This paper proposes a hybrid decision algorithm which can reduce unimportant block sizes and prediction modes. Entropy feature of each MB is extracted to decide the optimal block size. Then Prewitt operators based edge direction extraction module is designed to select only promising prediction modes. In addition, above two decision modules can function together to form a fast hybrid decision algorithm for intra prediction. Sufficient experiments prove that this hybrid algorithm is able to achieve 63{\%} time reduction with PSNR loss limited to 0.07dB, bit rate increase at most 2.02{\%}. It is expected to be utilized as a favorable accelerator for a real-time H.264 encoder for HDTV.",
keywords = "Block Size Decision, Entropy Theory, H.264/AVC, Intra Prediction, Mode Decision",
author = "Guifen Tian and Tianruo Zhang and Takeshi Ikenaga and Satoshi Goto",
year = "2009",
doi = "10.1007/978-3-540-92892-8_10",
language = "English",
isbn = "354092891X",
volume = "5371 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "85--95",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - A fast hybrid decision algorithm for H.264/AVC intra prediction based on entropy theory

AU - Tian, Guifen

AU - Zhang, Tianruo

AU - Ikenaga, Takeshi

AU - Goto, Satoshi

PY - 2009

Y1 - 2009

N2 - Rate Distortion Optimization based spatial intra coding is a new feature of H.264/AVC standard. It efficiently improves the video coding performance by brutally utilizing variable block sizes and multiple prediction modes. Thus, extremely high computation complexity is required. This paper proposes a hybrid decision algorithm which can reduce unimportant block sizes and prediction modes. Entropy feature of each MB is extracted to decide the optimal block size. Then Prewitt operators based edge direction extraction module is designed to select only promising prediction modes. In addition, above two decision modules can function together to form a fast hybrid decision algorithm for intra prediction. Sufficient experiments prove that this hybrid algorithm is able to achieve 63% time reduction with PSNR loss limited to 0.07dB, bit rate increase at most 2.02%. It is expected to be utilized as a favorable accelerator for a real-time H.264 encoder for HDTV.

AB - Rate Distortion Optimization based spatial intra coding is a new feature of H.264/AVC standard. It efficiently improves the video coding performance by brutally utilizing variable block sizes and multiple prediction modes. Thus, extremely high computation complexity is required. This paper proposes a hybrid decision algorithm which can reduce unimportant block sizes and prediction modes. Entropy feature of each MB is extracted to decide the optimal block size. Then Prewitt operators based edge direction extraction module is designed to select only promising prediction modes. In addition, above two decision modules can function together to form a fast hybrid decision algorithm for intra prediction. Sufficient experiments prove that this hybrid algorithm is able to achieve 63% time reduction with PSNR loss limited to 0.07dB, bit rate increase at most 2.02%. It is expected to be utilized as a favorable accelerator for a real-time H.264 encoder for HDTV.

KW - Block Size Decision

KW - Entropy Theory

KW - H.264/AVC

KW - Intra Prediction

KW - Mode Decision

UR - http://www.scopus.com/inward/record.url?scp=59149093225&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=59149093225&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-92892-8_10

DO - 10.1007/978-3-540-92892-8_10

M3 - Conference contribution

AN - SCOPUS:59149093225

SN - 354092891X

SN - 9783540928911

VL - 5371 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 85

EP - 95

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -