Measurements of individual color matching functions to solve the discrepancy in color appearance among color display devices

本研究的目的是通过假设可用于计算三刺激值的颜色匹配函数具有显著的个体差异，来处理在不同颜色设备之间实现同色异谱匹配时的三刺激值不匹配的问题。为了实现我们的目标，我们首先需要开发一种紧凑的仪器来精确地测量颜色匹配函数，以及用我们的仪器估计颜色匹配函数的方法。然后我们需要验证我们的方法，然后收集许多观察者的数据，以澄清个体差异和与CIE标准观察者的偏差。我们开发了一种设备，可以用最大视点系统呈现单色光的任意组合。我们还开发了一种代数方法，使用M系列将独立的单色光集合呈现给观察者，以获得颜色匹配函数。通过将该方法与传统的最大饱和度方法进行比较，结果表明该方法的抗噪能力不如MSM方法。由于我们的方法是基于矩阵运算，特定刺激中的噪声穿透到其他部分，而在MSM中，噪声仅限于该刺激。另一方面，当参考刺激在整个实验过程中保持不变时，该方法的匹配满意度要高得多。10名受试者作为观察者参与了我们的实验，估计的CMF不仅在B CMF上表现出个体差异，而且在G和R CMF上也表现出个体差异。我们还进行了色纸和CRT显示器之间的同色异谱配色实验，并试图补偿CMF的个体差异以获得定制的三刺激值。两种不同介质的三刺激值之间的色差或欧几里德距离显示出定制的三刺激值比标准观察者计算的值小的趋势。

The purpose of this study is to deal with the problem of the mismatch of the tristimulus value when the metameric match is achieved among different color devices, by assuming that the color matching functions, which can be used to calculate tristimulus value, have significant individual differences. In order to achieve our goals, we firstly need to develop a compact apparatus to measure color matching functions precisely, and a method of estimating color matching functions with our apparatus. Then we need to verify our method, followed by collecting data on many observers to clarify the individual differences and that deviation from CIE Standard Observer.We developed an apparatus that can present arbitrary combinations of the monochromatic lights with a Maximillian view system. Also we developed an algebraic method using M series to present independent sets of monochromatic lights to the observer to obtain color matching functions. By comparing our method with a traditional method, maximum saturation method, it turned out that our method is less noise-tolerant than MSM. As our method is based on the matrix operation, the noise in a specific stimulus penetrates to other parts, while in the MSM the noise is restricted to that stimulus. The satisfaction of the match, on the other hand, is much higher in our method as the reference stimulus remained constant through all the experiment.10 subjects participated in our experiment as an observer, and the estimated CMFs showed individual differences, not only in the B CMFs but also in G and R CMFs.We also conducted metameric color matching experiment between a color paper and a CRT monitor, and tried to compensate individual differences of CMFs to obtain customized tristimulus values. The color differences, or Eucrid distance, between the tristimulus values of two different media showed a trend that customized tristimulus values give smaller values than those calculated from Standard observers.