Scholarlink

<p>In the process of color reproduction, accurate prediction of the color characteristics of color halftone images and the establishment of a color prediction model for the spectral reflectance of halftone images is an issue of great concern in the field of print image device characterization and color print image quality control. Traditional models such as Murray-Davis, Clapper-Yule, Yule-Nielsen and modified Yule-Nielsen spectral Neugebauer models are widely used in this field due to their excellent color and spectral prediction accuracy. However, current research shows that none of these models takes into account the use of black ink in the printing process, which is a significant limitation in predicting the spectral properties of CMYK four-color inks. Additionally, these models have relatively large errors in predicting light color tones. In this study, we employ Neugebauer’s equation to extend the conventional 4-color CMYK input to 16-color inputs with a least-squares support vector machine (LSSVM) model optimized by an improved particle swarm optimization (IPSO) algorithm. The spectral reflectance output is efficiently reduced and restored by a PCA-based principal component extraction algorithm, which ensures a high degree of accuracy of the spectral reflectance in the dataset, with a principal component contribution of up to 99.99%. The experimental results show that the proposed model has significantly lower RMSE values and CIEDE2000 color differences compared to existing methods. Thus, it can effectively reduce the color differences in the printing process, greatly reduce material and energy consumption, and significantly improve the quality of printed images.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Color spectral reflectance prediction model based on machine learning

  • 期刊论文

  • 发表日期：

    2026年6月16日

查看全文
  • Dongwen Tian,
  • Jinghuan Ge,
  • Na Su,
  • Qian Cao

摘要

In the process of color reproduction, accurate prediction of the color characteristics of color halftone images and the establishment of a color prediction model for the spectral reflectance of halftone images is an issue of great concern in the field of print image device characterization and color print image quality control. Traditional models such as Murray-Davis, Clapper-Yule, Yule-Nielsen and modified Yule-Nielsen spectral Neugebauer models are widely used in this field due to their excellent color and spectral prediction accuracy. However, current research shows that none of these models takes into account the use of black ink in the printing process, which is a significant limitation in predicting the spectral properties of CMYK four-color inks. Additionally, these models have relatively large errors in predicting light color tones. In this study, we employ Neugebauer’s equation to extend the conventional 4-color CMYK input to 16-color inputs with a least-squares support vector machine (LSSVM) model optimized by an improved particle swarm optimization (IPSO) algorithm. The spectral reflectance output is efficiently reduced and restored by a PCA-based principal component extraction algorithm, which ensures a high degree of accuracy of the spectral reflectance in the dataset, with a principal component contribution of up to 99.99%. The experimental results show that the proposed model has significantly lower RMSE values and CIEDE2000 color differences compared to existing methods. Thus, it can effectively reduce the color differences in the printing process, greatly reduce material and energy consumption, and significantly improve the quality of printed images.