Temporal video transcoding based on frame complexity analysis for mobile video communication

Chia Hung Yeh*, Shu Jhen Fan Jiang, Chih Yang Lin, Mei Juan Chen

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Mobile communication systems provide more kinds of advanced video services. One of the most significant problems of mobile video communication is to adapt video content to different network characteristics, which is critical to acceptable visual quality of the transmitted videos. This paper proposes a frame selection scheme for temporal video transcoding through frame complexity analysis that consists of visual complexity and temporal coherence. Visual complexity represents the importance of frames and temporal coherence reveals the consequences of frame skipping; both of these two features are good indicators to determine if a frame should be kept. Experimental results show that our proposed method achieves higher visual quality and has more encoded frames than other existing schemes in the same bitrate, which means that our proposed method achieves high accuracy in selecting the representative frames for temporal video transcoding. The proposed method also has better video quality assessment scores than that of other existing methods, which means the proposed method keeps higher level of integrity of the transcoded videos.

Original languageEnglish
Article number6324395
Pages (from-to)38-46
Number of pages9
JournalIEEE Transactions on Broadcasting
Volume59
Issue number1
DOIs
Publication statusPublished - 2013
Externally publishedYes

Keywords

  • Frame complexity
  • frame selection
  • temporal coherence
  • temporal transcoding
  • video quality assessment
  • video transcoding
  • visual complexity

ASJC Scopus subject areas

  • Media Technology
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Temporal video transcoding based on frame complexity analysis for mobile video communication'. Together they form a unique fingerprint.

Cite this