• Laser & Optoelectronics Progress
  • Vol. 57, Issue 1, 013003 (2020)
Qiulan Yang, Xiaoxia Wan*, and Gensheng Xiao
Author Affiliations
  • Department of Printing and Packaging, Wuhan University, Wuhan, Hubei 430072, China
  • show less
    DOI: 10.3788/LOP57.013003 Cite this Article Set citation alerts
    Qiulan Yang, Xiaoxia Wan, Gensheng Xiao. Multispectral Dimension Reduction Algorithm Based on Partial Least Squares[J]. Laser & Optoelectronics Progress, 2020, 57(1): 013003 Copy Citation Text show less
    Chromaticity diagrams of 1600 Munsell color blocks obtained by CIE D65 and CIE 1931 standard observers. (a) L*-a* chromaticity diagram; (b) L*-b* chromaticity diagram; (c) a*-b* chromaticity diagram
    Fig. 1. Chromaticity diagrams of 1600 Munsell color blocks obtained by CIE D65 and CIE 1931 standard observers. (a) L*-a* chromaticity diagram; (b) L*-b* chromaticity diagram; (c) a*-b* chromaticity diagram
    Spectral error distributions obtained by different dimension reduction methods. (a) Dimension reduction by LabPQR method; (b) dimension reduction by LabKMN method
    Fig. 2. Spectral error distributions obtained by different dimension reduction methods. (a) Dimension reduction by LabPQR method; (b) dimension reduction by LabKMN method
    Spectral reconstruction RMSE curves obtained by different dimension reduction methods. (a) Dimension reduction by LabPQR method; (b) dimension reduction by LabKMN method
    Fig. 3. Spectral reconstruction RMSE curves obtained by different dimension reduction methods. (a) Dimension reduction by LabPQR method; (b) dimension reduction by LabKMN method
    Fitted curves of reconstructed spectra for different samples. (a) Sample with good fitted effect; (b) sample with general fitted effect; (c) sample with poor fitted effect
    Fig. 4. Fitted curves of reconstructed spectra for different samples. (a) Sample with good fitted effect; (b) sample with general fitted effect; (c) sample with poor fitted effect
    Dimensionreduction methodRMSEΔEab*(D65)ΔEab* (A)Sampleeligibility rate /%
    MaxMinMeanMaxMinMeanMaxMinMean
    LabPQR0.11780.00240.01648.39300.18652.921410.59460.19492.819784.37
    LabKMN0.07700.00220.01395.72200.09801.502511.74580.03832.125195.28
    Table 1. Comparison of spectral reconstruction accuracy between LabPQR dimension reduction method and LabKMN dimension reduction method
    Qiulan Yang, Xiaoxia Wan, Gensheng Xiao. Multispectral Dimension Reduction Algorithm Based on Partial Least Squares[J]. Laser & Optoelectronics Progress, 2020, 57(1): 013003
    Download Citation