基于图像质量度量的纸基全彩3D打印颜色再现评价

Color Reproduction Evaluation of Paper-Based Full-Color 3D Printing Based on Image Quality Metrics

全彩3D打印能为三维数字制造提供真实的颜色再现及控制,但由于缺乏客观可靠的颜色再现质量评价工具,使其在个性化定制中一直面临棘手的颜色精确再现难题。本研究针对具有环保及低成本优势的纸基全彩3D打印技术,利用设计的特定三维色标模型,对高清相机获取的主视面图像进行平均结构相似度(MSSIM)、特征相似度(FSIMc)以及视觉差异度(iCID)等图像基度量的计算及统计,再以客观测量的色差进行量化关联分析。实验结果表明,当前测试样的图像基度量与色差线性相关,但对应线性拟合方程还需要通过更多测试样进行参数优化。其中,彩色SSIM映射图能提供纸基3D打印模型的颜色再现细节评价,但上述三个图像基度量在纸基全彩3D测试模型颜色再现评价中表现了相对独立性。该研究能为更符合人眼视觉感知的全彩3D打印颜色再现评价的图像基度量开发提供理论参考。

Full-color 3D printing can provide realistic color reproduction and control for 3D digital fabrication,but the lack of objective and reliable tools for evaluating color quality has always led to the thorny issue of color reproduction in customized applications.In this study,the paper-based full-color 3D printing technology with its environmental and low-cost advantages was taken as an example.Combining the designed 3D color scaling model and the key visual images acquired from a high-definition camera,objective metrics such as mean-structural similarity(MSSIM),color feature similarity(FSIMc)and improved color image difference(iCID)were calculated and counted,and then quantified correlation analysis was performed on the measured color differences.The experimental results showed that the image quality metrics are well correlated with color difference,but its unified linear correlation requires to optimize with more tested samples.Moreover,structural similarity is better than feature similarity in evaluating color reproduction quality of paper-based 3D printing models,while color SSIM mappings provide more details for color reproduction of paperbased full-color 3D printed models,but the above three image quality metrics demonstrate relative independence in color reproduction evaluation of paper-based 3D test models.This study can provide a theoretical reference for the development of other image quality metrics for color reproduction evaluation of full-color 3D printing that is more consistent with the human eye’s visual perception.