Improved scanning methods for wavelet coefficients of video signals

Fabrice Rossetti, Jiro Katto, Mutsumi Ohta

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

This paper describes novel methods to improve coding efficiency of the wavelet transform from the viewpoint of scanning techniques. Firstly, statistical properties of the wavelet transform are examined using real image sequences. They indicate energy distribution or correlations of wavelet transformed coefficients. Taking them into account, efficient scanning methods are investigated for two different approaches. The first approach treats transform coefficients, located in spatially neighboring position, in 8 × 8 blocks. This is called block based approach, and its optimum scanning path is examined similar to the zigzag scan for DCT (discrete cosine transform). Furthermore, a new scanning algorithm utilizing correlation between transform coefficients is proposed. The second approach carries out scanning over the entire subimage created by the wavelet transform. This is called subband based approach, in which scanning is applied to each subband. A new scanning algorithm is presented utilizing a block-wise non-linear filter which removes isolated non-zero coefficients. Simulations are carried out, and their results show the effectiveness of the above approaches especially for low bit-rate applications.

Original languageEnglish
Pages (from-to)365-378
Number of pages14
JournalSignal Processing: Image Communication
Volume8
Issue number4
DOIs
Publication statusPublished - 1996 May
Externally publishedYes

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Improved scanning methods for wavelet coefficients of video signals'. Together they form a unique fingerprint.

  • Cite this