The clustered causal state algorithm efficient pattern discovery for lossy data-compression applications

Mendel Schmiedekamp, Aparna Subbu, Shashi Phoha

Research output: Contribution to journalReview articlepeer-review

8 Scopus citations

Abstract

Clustered Causal State Algorithm (CCSA), a pattern discovery algorithm, is developed in lossy video compression to approximate E-machines, for use in real time and resource limited applications. CCSA performs unsupervised pattern discovery, producing pattern descriptions with computational efficiency for use in data compression in exchange for a small loss in description fidelity. It is based on the hierarchical agglomerative clustering method and attempts to describe patterns intrinsic to a process, which it achieves at a lower computational cost. The inputs to the CCSA program are the symbol stream and the algorithm executes in the following steps: initialization, clustering, finalization. CCSA has the distinct advantage of polynomial computational complexity, and using this algorithm image compression takes few seconds and it could reliably generate 10 to 20 fold compressions.

Original languageEnglish (US)
Article number1677484
Pages (from-to)59-67
Number of pages9
JournalComputing in Science and Engineering
Volume8
Issue number5
DOIs
StatePublished - Sep 2006

All Science Journal Classification (ASJC) codes

  • General Computer Science
  • General Engineering

Fingerprint

Dive into the research topics of 'The clustered causal state algorithm efficient pattern discovery for lossy data-compression applications'. Together they form a unique fingerprint.

Cite this