Fast QTMT partition decision based on deep learning

Compared with the predecessor standards, versatile video coding (VVC) significantly improves compression efficiency by a quadtree with nested multi-type tree (QTMT) structure but at the expense of extremely high coding complexity.To reduce the coding complexity of VVC, a fast QTMT partition method w...

Full description

Saved in:
Bibliographic Details
Main Authors: Shuang PENG, Xiaodong WANG, Zongju PENG, Fen CHEN
Format: Article
Language:zho
Published: Beijing Xintong Media Co., Ltd 2021-04-01
Series:Dianxin kexue
Subjects:
Online Access:http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2021062/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1841529133007372288
author Shuang PENG
Xiaodong WANG
Zongju PENG
Fen CHEN
author_facet Shuang PENG
Xiaodong WANG
Zongju PENG
Fen CHEN
author_sort Shuang PENG
collection DOAJ
description Compared with the predecessor standards, versatile video coding (VVC) significantly improves compression efficiency by a quadtree with nested multi-type tree (QTMT) structure but at the expense of extremely high coding complexity.To reduce the coding complexity of VVC, a fast QTMT partition method was proposed based on deep learning.Firstly, an attention-asymmetric convolutional neural network was proposed to predict the probability of partition modes.Then, the fast decision of partition modes based on the threshold was proposed.Finally, the cost of coding performance and time was proposed to obtain the optimal threshold, and the threshold decision method was proposed.Experimental results at different levels show that the proposed method achieves an average time saving of 48.62%/52.93%/62.01% with the negligible BDBR of 1.05%/1.33%/2.38%.Such results demonstrate that the proposed method significantly outperforms other state-of-the-art methods.
format Article
id doaj-art-ec870f4f626a43cf9a7ae0b40c559bd1
institution Kabale University
issn 1000-0801
language zho
publishDate 2021-04-01
publisher Beijing Xintong Media Co., Ltd
record_format Article
series Dianxin kexue
spelling doaj-art-ec870f4f626a43cf9a7ae0b40c559bd12025-01-15T03:26:07ZzhoBeijing Xintong Media Co., LtdDianxin kexue1000-08012021-04-0137738159807627Fast QTMT partition decision based on deep learningShuang PENGXiaodong WANGZongju PENGFen CHENCompared with the predecessor standards, versatile video coding (VVC) significantly improves compression efficiency by a quadtree with nested multi-type tree (QTMT) structure but at the expense of extremely high coding complexity.To reduce the coding complexity of VVC, a fast QTMT partition method was proposed based on deep learning.Firstly, an attention-asymmetric convolutional neural network was proposed to predict the probability of partition modes.Then, the fast decision of partition modes based on the threshold was proposed.Finally, the cost of coding performance and time was proposed to obtain the optimal threshold, and the threshold decision method was proposed.Experimental results at different levels show that the proposed method achieves an average time saving of 48.62%/52.93%/62.01% with the negligible BDBR of 1.05%/1.33%/2.38%.Such results demonstrate that the proposed method significantly outperforms other state-of-the-art methods.http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2021062/VVCQTMTfast partition decisiondeep learning
spellingShingle Shuang PENG
Xiaodong WANG
Zongju PENG
Fen CHEN
Fast QTMT partition decision based on deep learning
Dianxin kexue
VVC
QTMT
fast partition decision
deep learning
title Fast QTMT partition decision based on deep learning
title_full Fast QTMT partition decision based on deep learning
title_fullStr Fast QTMT partition decision based on deep learning
title_full_unstemmed Fast QTMT partition decision based on deep learning
title_short Fast QTMT partition decision based on deep learning
title_sort fast qtmt partition decision based on deep learning
topic VVC
QTMT
fast partition decision
deep learning
url http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2021062/
work_keys_str_mv AT shuangpeng fastqtmtpartitiondecisionbasedondeeplearning
AT xiaodongwang fastqtmtpartitiondecisionbasedondeeplearning
AT zongjupeng fastqtmtpartitiondecisionbasedondeeplearning
AT fenchen fastqtmtpartitiondecisionbasedondeeplearning