Related Experiment Videos
Using reflectance models for color scanner calibration.
1Department of Electrical and Computer Engineering, University of California, Irvine, 92697, USA. mshi@ece.uci.edu
Summary
Linear spectral reflectance models improve color scanner calibration accuracy. Incorporating reflectance information enhances device-independent CIE XYZ value generation beyond standard polynomial regression methods.
Area of Science:
- Color Science
- Image Processing
- Metrology
Background:
- Color scanner calibration aims to convert scanner vectors to device-independent CIE XYZ values.
- Polynomial regression is a common but potentially inaccurate calibration method.
- Linear spectral reflectance models accurately describe many materials.
Purpose of the Study:
- To investigate the efficacy of linear spectral reflectance models for color scanner calibration.
- To compare the performance of reflectance-based calibration with standard polynomial regression.
- To determine if incorporating reflectance information improves CIE XYZ value accuracy.
Main Methods:
- Applying linear spectral reflectance models to scanner calibration.
- Viewing color scanner calibration as a reflectance estimation problem.
- Developing and evaluating methods that exploit reflectance function information.
Main Results:
- Linear models alone often insufficient for exact CIE XYZ recovery from scanner vectors.
- Exploiting reflectance information significantly improves calibration accuracy.
- Reflectance-based methods outperform standard polynomial regression on the same input data.
Conclusions:
- Linear spectral reflectance models offer a valuable framework for enhancing color scanner calibration.
- Integrating knowledge of material reflectance properties leads to more accurate device-independent color measurements.
- This approach provides a more robust alternative to purely data-driven regression techniques.