Research on optimal control for spatio-temporal video quality in scalable video coding

2014 Fiscal Year Final Research Report

• Project/Area Number: 24500112
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Media informatics/Database
• Research Institution: University of Fukui
• Principal Investigator: YOSHIDA Toshiyuki 福井大学, 工学(系)研究科(研究院), 教授 (50240297)
• Project Period (FY): 2012-04-01 – 2015-03-31
• Keywords: 動画像符号化 / 階層符号化 / 画質制御 / レート制御 / MOS

Outline of Final Research Achievements

A video quality control technique for scalable video coding has been proposed and enstablished in this research. The technique focuses on SNR-temporal scalable video coding, where SNR and temporal scalability gradually improves quality in terms of each video frame and motion smoothness, respectively, and maximizes an estimate for spatio-temporal mean opinion score (MOS) of each video scene in each scalable layer. A video quality equalization technique based on variable-bit-rate video coding has also been introduced in the proposed framework in order to reduce MOS fluctuation over scenes. Finally, total advantage of the proposed technique has been evalulated for several test sequences comprising multiple scenes.

Free Research Field

画像処理・信号処理