Premium
A New matching strategy: Trial of the principal component coordinates
Author(s) -
Agahian Farnaz,
Amirshahi Seyed Hossein
Publication year - 2008
Publication title -
color research and application
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.393
H-Index - 62
eISSN - 1520-6378
pISSN - 0361-2317
DOI - 10.1002/col.20364
Subject(s) - principal component analysis , eigenvalues and eigenvectors , mathematics , matching (statistics) , sample (material) , basis (linear algebra) , color space , pattern recognition (psychology) , population , artificial intelligence , algorithm , computer science , statistics , geometry , image (mathematics) , chemistry , demography , physics , chromatography , quantum mechanics , sociology
A new matching strategy based on the equalization of the first three principal component coordinates of sample and target in a 3D eigenvector space is stated. Two series of databases including 1269 specimens of Munsell Color Book and a virtual sample population of textile materials were selected. Their first three basis functions were extracted and considered as axes of eigenvector space. The principal component coordinates of two different collections of textile samples were determined in these spaces and considered as matching criteria. The performance of the proposed algorithm is evaluated by the color difference values under different light sources as well as the root mean square differences of reflectance curves. Results indicate some types of improvements in comparison with previous algorithms in terms of spectral as well as colorimetric accuracy. © 2007 Wiley Periodicals, Inc. Col Res Appl, 33, 10–18, 2008