An evolvable predictor for lossless image compression

Daniel León, Sina Balkir, Khalid Sayood

Research output: Contribution to journalConference article


This paper presents the adaptation via evolutionary techniques of a pixel predictor for lossless image compression. The pixel prediction is based on a linear combination of some neighbor pixels. The evolutionary algorithm selects the coefficients and the pixels involved in the pixel prediction. Experiments carried out on gray level images of the proposed system show a performance comparable and in some cases better than existing predictive coding techniques.

Original languageEnglish (US)
Pages (from-to)IV/731-IV/734
JournalProceedings - IEEE International Symposium on Circuits and Systems
Publication statusPublished - Jan 1 2002
Event2002 IEEE International Symposium on Circuits and Systems - Phoenix, AZ, United States
Duration: May 26 2002May 29 2002


ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this