Video Coding Using Dissolve-Compensated Prediction

Hirohisa Jozawa*, Kazuto Kamikura, Hiroshi Watanabe, Hiroshi Kotera

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

This paper describes a novel motion compensation method for dissolve images where the current shot gradually disappears and the next shot appears instead. The conventional MPEG-like prediction method does not perform well for dissolve images since two shots are mixed. To solve this problem, a new bi-directional prediction method and a motion estimation method suitable for dissolve images are proposed as well as a dissolve detection method. Experimental results showed that the dissolve compensation significantly outperforms the MPEG' B-picture in terms of prediction efficiency. By using the proposed method, the mean-square errors (MSE) are reduced to 1/4 of those of the conventional B-pictures.

Original languageEnglish
Pages (from-to)X52-236
JournalITG-Fachbericht
Issue number143
Publication statusPublished - 1997 Dec 1
Externally publishedYes

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture

Fingerprint

Dive into the research topics of 'Video Coding Using Dissolve-Compensated Prediction'. Together they form a unique fingerprint.

Cite this