An effective colour feature extraction method using evolutionary computation for face recognition. (1st January 2011)

• Record Type: Journal Article
• Authors: Shih, Peichung; Liu, Chengjun
• Is Part Of: International journal of biometrics. Volume 3:Number 3(2011)
• Page Start: 206
• Page End: 227
• Publication Date: 2011-01-01
• Subjects: BEE -- biometric experimentation environment -- colour space -- GAs -- genetic algorithms -- FRGC -- face recognition grand challenge -- gallery -- probe -- query -- target; Biometric identification -- Periodicals; Biometry -- Periodicals; 570.15195
• DOI: 10.1504/IJBM.2011.040816
• Languages: English
• ISSNs: 1755-8301

Abstract

This paper presents a colour feature extraction method using a Genetic Algorithm (GA) to seek the optimal colour space transformation that leads to an effective image representation for face recognition. A new colour space, LC1C2, consisting of one luminance (L) channel and two chrominance channels (C1, C2) is introduced as a linear transformation of the input RGB colour space. The specific transformation from the RGB colour space to the LC1C2 colour space is optimised by a GA where a fitness function guides the evolution towards higher recognition accuracy.