"基于三维打印球体的色差计算研究:CIE公式适用性及优化效果"

版权申诉
0 下载量 33 浏览量 更新于2024-02-25 收藏 1.01MB DOCX 举报
The research focused on evaluating the suitability of existing International Commission on Illumination (CIE) color difference formulas (based on flat colors) for assessing color difference in three-dimensional (3D) printed sphere models. Using the Sailner J400 color 3D printer, 200 pairs of 4 cm diameter sphere models in five color centers recommended by CIE (gray, red, yellow, green, blue) were created, with 150 pairs in matte finish and 50 pairs in glossy finish. A total of 59 observers with normal color vision aged 19-26 years were organized to conduct color difference evaluation experiments using the grey scale method. The performance of three color spaces (CIELAB, CIECAM02, and CAM16) and eight color difference formulas were examined, and the power exponent and lightness factor of the color difference formulas were optimized. The results showed that the original color difference formulas had relatively consistent predictive performance for 3D sphere color difference, with standardized residual sum of squares (STRESS) values ranging from 32.5 to 34.7. The predictive performance for glossy sphere samples was better than matte sphere samples. F-tests indicated that the power exponent color difference calculation method established for 3D sphere samples significantly improved the performance of the original color difference formulas, with optimized STRESS values reduced to 24.5-27.3. However, the improvement after lightness factor optimization was not significant. In conclusion, the research provides insights into the evaluation of color difference in 3D printed sphere models using different color spaces and formulas, as well as the optimization of parameters for improved predictive performance. The findings contribute to the development of more accurate and reliable methods for assessing color difference in 3D printed objects, which is crucial for various applications in industries such as design, manufacturing, and quality control.