Representing Spectral Data Using Lab PQR Color Space in Comparison ‎with PCA Method

In many applications of color ‎technology such as spectral color ‎reproduction, it is of interest to ‎represent the spectral data with ‎lower dimensions than spectral ‎space dimensions. It is more than ‎half of a century that Principal ‎Component Analysis (PCA) ‎method has been applied to find ‎the number of independent basis ‎vectors of spectral dataset and ‎representing spectral reflectance ‎with lower dimensions. Recently, a ‎new Interim Connection Space ‎‎(ICS) named LabPQR was ‎introduced, which contains three ‎colorimetric dimensions and ‎additional black metamer space. In ‎the present study, the performance ‎of PCA method in comparison to ‎LabPQR was investigated for ‎representation of spectral ‎reflectance. For this end, different ‎color data sets including Munsell, ‎Glossy Munsell, ‎GretagMacbethColorChecker, ‎Esser test chart and two printing ‎datasets were evaluated. The ‎results show that, the performance ‎of PCA and LabPQR, depends on ‎the applied dataset. Based on ‎spectral metrics such as RMS and ‎GFC values, PCA has better ‎results than LabPQR. Considering ‎color difference errors, LabPQR is ‎a better space than PCA even ‎based on the color difference ‎under second illuminant. ‎Moreover, the dataset used for ‎obtaining PQR vectors affects the ‎representation results. For some ‎datasets, the PQR components of ‎the other sets perform better. ‎However, obtaining PQR bases ‎from the same data source, gives ‎better results. It was found that ‎Cohen and Kappauf-based ‎methods performs better for all the ‎datasets compared with ‎unconstrained LabPQR method. ‎