基于模式识别技术的色貌模型评价研究  

The research of color appearance models evaluation based on pattern recognition technology

作　　者：王晓红[1] 孙平[1] 徐卓[1] 吕兆锋[1] 

机构地区：[1]上海理工大学,上海200093

出　　处：《光学技术》2012年第5期573-578,共6页Optical Technique

摘　　要：选用Munsell新标数据集,采用基于模式识别技术的圆度、色相角偏差、明度线性度与空间投影点聚集度四个指标来分别评价目前八个典型色貌模型在彩度均匀性、色相预测准确性、明度均匀性、颜色的空间再现的能力。结果表明,不同的色貌模型在四个方面的颜色再现性上表现出不同的优势。八个色貌模型对彩度预测最好的是RLAB,每个色貌模型对红色和绿色预测比较好,CAM02在色相预测整体上占优势;明度的均匀性预测都比较好,尤其是CIELAB;LLAB的颜色的空间再现能力最好。Munsell new standard data set and four pattern recognition indexes , which are roundness, the deviation of hue angles, lightness linearity and gathered degrees of space projection points, are chosen to evaluate the current eight typical color appearance models in chroma uniformity, hue forecasting accuracy , the lightness uniformity and the repro- duction ability of color space. The result is that different color appearance models have different advantages in the four as- pects. The RLAB is good at the chroma predetermination. The red and green are predicted well by every color appearance and CAM02 has an advantage in hue reproduction completely. Every color appearance excels in the lightness uniformity, especially CIELAB. The color space is restored best by LLAB.

关 键 词：模式识别技术 色貌模型 孟塞尔新标数据集 圆度 色相角偏差 明度线性度 投影点聚集度 

分 类 号：O432.3[机械工程—光学工程]