Robustified Least Squares for Spectrum Neugebauer Model
HAN Fang-fang, LI Li-xia, ZHANG Yi-xin
Author information+
Jiangnan University, Wuxi 214122, China
Show less
History+
Issue Date
2016-11-21
Abstract
Yule-Nielsen modified spectrum Neugebauer model is an important spectrum prediction model in predicting color output of color halftone prints. In order to improve the prediction accuracy, spectrum Neugebauer model and robustified estimation theory were analyzed and a method of robustified least squares to estimate dot area function and primary spectrum reflectance were put forward. Commonly used robust estimation programs are Huber estimation and IGG estimation. Experiment results showed that compared to traditional least squares (LS) based methods, the accuracy of spectrum Neugebauer model with RLS approach is higher and more stable under the condition of gross error.