Reflectance spectra reconstruction from trichromatic camera based on kernel partial least square method.

Area of Science:

• Color Science
• Image Processing
• Computational Imaging

Background:

• Accurate spectral reflectance reconstruction is crucial for color management.
• Existing methods face challenges with trichromatic cameras and complex spectral data.
• Multispectral imaging offers rich color information but requires sophisticated analysis.

Purpose of the Study:

• To propose a novel spectral reflectance reconstruction method using kernel partial least square (KPLS) regression.
• To estimate spectral reflectance from a 9-channel multispectral imaging system utilizing a commercial trichromatic camera.
• To evaluate the performance of the proposed KPLS method against existing techniques.

Main Methods:

• Developed a spectral reconstruction method integrating partial least square (PLS) algorithm with kernel functions.
• Employed a 9-channel multispectral imaging system with a standard trichromatic camera.
• Validated the method using simulated and real camera responses from Munsell Matte color and IT8.7/3 datasets.

Main Results:

• The proposed KPLS method demonstrated superior or equivalent performance compared to existing spectral reconstruction techniques.
• Experimental results confirmed the method's effectiveness in estimating spectral reflectance.
• The method proved satisfactory for practical color management applications.

Conclusions:

• The novel KPLS-based spectral reflectance reconstruction method is effective and reliable.
• This approach enhances the capability of using trichromatic cameras for multispectral imaging applications.
• The method provides a robust solution for accurate color management.