Color pattern recognition with CIELAB coordinates

Ramon Corbalan Yuste, Josefa Yzuel Gimenez, M. Corbalán, MS. Millán, MJ. Yzuel

Research output: Contribution to journalArticleResearchpeer-review

29 Citations (Scopus)

Abstract

A color pattern recognition system must identify a target by its shape and color distribution. In real situations, however, the color information is affected by changes of the light source (e.g., from indoor illumination to outdoor daylight), often making recognition impossible. In this work, we propose a color pattern recognition technique with tolerance for illumination changes within the common sources of white light. This can be accomplished using the coordinates of the CIELAB system, luminance (L*), chroma (C*), and hue (h*) instead of the conventional RGB system. The proposal has some additional advantages: there is no need to store a matched filters base to analyze scenes captured under different light sources (one set of filters for each illuminant light source) and therefore the recognition process can be simplified; and in most cases, the contribution of only two channels (C* and h*) is enough to avoid false alarms in color pattern recognition. From the results, we show that the recognition system is improved when CIELAB coordinates are used. © 2002 Society of Photo-Optical Instrumentation Engineers.
Original languageEnglish
Pages (from-to)130-138
JournalOptical Engineering
Volume41
Issue number1
DOIs
Publication statusPublished - 1 Jan 2002

Keywords

  • CIELAB coordinates
  • Color CCD camera
  • Color images
  • Pattern recognition
  • White light sources

Fingerprint Dive into the research topics of 'Color pattern recognition with CIELAB coordinates'. Together they form a unique fingerprint.

Cite this