打印机和图像色域提取算法

针对当前色彩管理研究缺乏提取图像色域算法及可视化技术支持的现状，提出了一种打印机色域提取算法。先确定打印机色域边界点，通过这些边界点利用几何方法快速构造色域外壳，对设备色域进行精确描述，并对色域及其剖面进行真彩色可视化。利用对色域的剖分操作，构造基于剖面的色域匹配模型。     

面向专色印刷的色域匹配算法

针对专色印刷色域与图像色域不匹配的问题,提出一种色域映射算法.首先分析专色色域的几何特征,发现常规色域匹配方法不适用;进而对明度、色调和饱和度三属性分而治之,采用非线性压缩、灰度轴模拟和彩度裁剪实现色域匹配.实验结果表明,该算法获得的图像匹配结果在颜色连续性和视觉效果方面均有明显的提高.     

分区最大化色域边界描述器的空区插值算法

针对分区最大化色域边界描述器在计算图像色域边界时所产生的空缺分区问题,提出了一种基于相邻分区加权平均的色域空区插值算法。将空缺分区的边界点置于该分区的中心,求出边界点的球坐标角度;利用邻域加权平均法插值计算出该分区的球坐标半径。在线色域边界计算中,针对相邻分区边界点连线与等色相面重合所导致的求交问题,将交点设定为其中一个分区的色域边界点,避免了线色域边界和色域映射计算中出现的色域边界缺失问题。在评价部分,借助等比例压缩的色域映射方式,设计心理物理学实验评价了色域空区插值算法的效果,同时测试了算法的计算效率。评价结果显示：提出的色域空区插值算法能够有效提升色域边界描述器的精度,改善色域映射效果,而且不影响色域边界的计算效率。     

一种基于图像空间频率特征的色域匹配算法*

色域匹配是色彩管理的一项关键技术。通过对彩色图像的特征描述的分析,提出了一种与图像空间特征相关的自适应色域匹配算法。该算法通过将图像增强技术与传统的色域匹配算法相结合,克服了传统算法不能同时保持明度对比度对比度和彩度的缺点,有效地提高了图像的再现效果和精度。     

打印机色域可视化

针对色彩管理应用需求，给出一种打印机最大色域的生成，提取及可视化方法，通过优化CMYK四色墨水组合得到最大色域样本集，利用凸壳算法构造其几何模型，可对用几何模型描述的色域进行剖面操作，最后对色域及其剖面进行真彩可视化，基于上述技术的可视化软件已在VisualC 6.0下实现，可用于设备彩色特性，分色方法和色域映射算法的辅助研究。     

基于TMS320C6416的色域扩展方法

由于激光电视覆盖的颜色区域是传统电视的颜色区域1.98倍,所以在激光电视等领域中需要一种转换系统快速完成荧光色域到激光色域的颜色扩展。本文给出了三组不同波长激光的色域扩展矩阵,在基于TMS320C6416的图像处理平台上实现了色域的转换,并且利用多种加速算法,加快了处理速度。实验结果表明,本文的方法有效地扩展了激光电视的颜色区域,降低了图像的饱和度。     

一种保肤色的压缩类色域匹配方法

肤色是一种重要的习惯色，打印时应尽量保其不变。文章在实验提取肤色域几何特征的基础上，通过对比分析肤色域与打印机色域的几何关系，提出一种保肤色的压缩类色域匹配算法。该算法以肤色域和打印机色域的分布知识为基础，随色调连续调整内核系数，可保大部分肤色不变，对包含肤色的图像有更理想的匹配效果。     

一种面向打印的自适应色域匹配方法

基于打印机色域的几何特征 ,提出一种按彩度最大点定锚点的裁剪类色域匹配算法。与现有同类方法相比 ,算法具有自适应性 ,可有效减小最坏子域的匹配失真。     

基于颜色与结构失真的色域映射图像无参考质量评价算法

为了预测色域映射图像客观质量,通过分析不同色域映射算法的映射原理发现色域映射图像中主要存在颜色失真与结构失真的情况.基于此,提出了一种基于颜色与结构失真的色域映射图像无参考质量评价算法.在颜色失真方面,计算色调异常率和图像R、G、B三个分量的统计分布与理想均匀分布之间的相对熵;在结构退化方面,提取图像的信息熵与四阶矩,并对图像亮度与饱和度进行统计建模,提取参数特征.随后,将以上提取的数据作为质量感知特征与图像的主观分数值输入后向传播神经网络进行回归训练得到针对色域映射图像的质量评价模型.最后,在三个公开的色域映射图像数据库上进行性能验证.实验结果表明,该算法在预测色域映射图像质量方面优于现有的无参考算法.     

带自由曲面的三视图重建算法研究

针对蕴含自由曲面的三视图,提出了通过恢复边界线和轮廓线进而根据这些恢复 的空间曲线重建自由曲面的算法。首先分析并证明了自由曲面在三视图中的投影性质,从而提 出边界投影和轮廓投影的匹配算法。针对视图中存在的被打断的样条曲线,提出了分段样条曲 线的爬坡算法来解决此类曲线不能匹配的问题。然后基于投影匹配序列重建出自由曲面的边界 线,再由边界投影上的点和轮廓投影端点的对应关系重建出空间轮廓线。由轮廓线等参采样构 造截面线并和边界线一起蒙皮生成最终自由曲面。本文提出的算法扩展了工程图的重建域。     

Measuring the color capability of modern display systems

Historically displays used three colorants in an additive system. During that time, the CIE chromaticity diagram adequately illustrated color capability. Modern displays are not constrained by this additive architecture, and the diagram can fail in its purpose. This is demonstrated by analysis and a large number of display measurements. A device‐independent methodology using CIE 1976 L*a*b* color gamut volume is described that provides a robust means to determine the size of the color gamut. This methodology is then extended to the ‘gamut rings’ diagram as a solution for visualizing color capability that directly correlates to color gamut volume. It is further shown how the methodology can be applied to determine the intersection between two gamut volume boundaries.     

Heuristic approach to selecting a color appearance model for hue‐preserved wide‐color gamut mapping

With the advent of wide‐gamut system colorimetry for ultra‐high definition television, the development of a gamut mapping algorithm for wide to standard gamuts is greatly needed. Most gamut mapping techniques have been developed to retain the perceived hues of the source as predicted by a color appearance model. However, some color appearance models erroneously predict the saturated colors, generating a serious hue discontinuity in wide‐color gamut mapping onto the Rec. 709 gamut boundary. This paper presents a heuristic approach to determining whether a color appearance model may create such artifacts when used in hue‐preserved gamut mapping algorithms.     

Analysis of gamut mapping algorithms from the viewpoint of color name matching

The validity of color name matching in analyzing gamut mapping algorithms is described in this paper. The performances of three types of gamut mapping algorithms (cusp scaling, clipping, and categorical weighting) are evaluated in a psychophysical experiment from the viewpoint of color name matching. Based on this evaluation, we have found that cusp‐scaling algorithm is capable of controlling lightness reasonably well, compensating for a decrease in chroma caused by the difference in the size of device gamut between a source and a destination device. Ranking of reproducible performance by the three gamut mapping algorithms in the psychophysical experiment is arrived at on the basis of color name matching.     

A method for selecting display primaries to match a target color gamut

Abstract— A method for selecting primaries of a wide‐gamut display is proposed, in which display color gamut is designed to match a target color gamut in CIELAB color space. A standard deviation of the relative maximum chroma of display and target color gamuts is defined. The selection method optimizes display primaries for the minimum standard deviation so that display and target color gamuts are similar in shape. It is shown that the color gamut of a laser display designed by this method is similar in shape to the theoretical maximum, or optimal, color gamut of objects. It is also shown that the color gamut of an LED display can be designed to include 99.7% of the gamut of Pointer's real‐world surface colors. LED primaries are selected to minimize the standard deviation of the relative maximum chroma of effective display color gamut and a target color gamut which is defined to include Pointer's real‐world surface colors. For both the laser and LED displays, it is necessary to constrain the red‐primary wavelength to avoid excessive optical power for the red primary.     

Preferred and maximally acceptable color gamut for reproducing natural image content

Abstract— With the development of wide‐gamut display technology, the need is clear for understanding the required size and shape of color gamut from the viewers' perspective. To that end, experiments were conducted to explore color‐gamut requirements based on viewers' preferred level of chroma enhancement of standard‐gamut images. Chroma preferences were measured for multiple hues using single‐hue images, and a corresponding hue‐dependent preferred chroma enhancement was successfully applied to natural, multi‐hue images. The multi‐hue images showed overall success, though viewers indicated that reds could be decreased even further in colorfulness, and yellows could be increased, which may argue in favor of multi‐primary displays. Viewer preferences do vary within the population, primarily in overall chroma level, and the differences can be largely accounted for with a single parameter for chroma‐level adjustment that includes the preferred hue dependence. Image content dependencies were also found, but they remain too complex to model. The hue‐dependent chroma preference results can be applied to display design and color‐enhancement algorithms.     

Effects of display properties on perceived color‐gamut volume and preference

Abstract— The effect of varying the color gamut of an extended‐gamut LCD on color appearance and preference was measured psychometrically in two experiments at each of two separate laboratories over a representative set of 10 images each. The first experiment measured the effect of color gamut on appearance, and the effect on the appearance attribute colorfulness was shown to be relatively strong compared with other attributes as the volume of display color gamut is varied. Overall, colorfulness monotonically increased at constant sensitivity as the gamut area in xy chromaticities increased while tending to become less and less sensitive to increasing the gamut volumes in CIELAB and CIECAM02. In the second experiment, the overall preference indicated an optimal color gamut for the display gamut volume even though the results were shown to be highly scene dependent.     

Power‐saving primary design for displays enclosing a target color gamut

Abstract— Display primaries are optimized for the trade‐off between the total primary power and color gamut under the requirement that a target color gamut is enclosed by the color gamut of the display. LED displays and HDTV color gamut are taken as examples. Compared to the display using a set of typical commercial RGB LEDs, it was found that a total optical (electrical) power of 23.6% (15.6%) can be saved for the display using optimal RGB LEDs. Although the size of the display color gamut is sacrificed, the color gamut of the display using optimal RGB LEDs still encloses the HDTV color gamut. The combined effect of the LED luminous efficiency and white‐point condition on the determination of the optimal LED wavelengths and bandwidths is also studied.     

A novel spectrum‐sequential display design with a wide color gamut and reduced color breakup

Abstract— The advantage of RGB color‐sequential displays is that they have no color filters, but the disadvantage is that they need to run at high refresh rates (> >180 Hz) to prevent flicker and color breakup. An alternative color‐sequential display, which can operate at relatively low refresh rates (～ 100 Hz) without disturbing color breakup or flicker, has been developed. The display has two color filters per pixel (cyan and magenta) on the LCD panel and the backlight can generate two types of spectra (blue‐green and green‐red), which results in a wide gamut four‐primary display, effectively. One part of the paper describes the color reproduction, including color‐filter design, gamut mapping, and multi‐primary conversion. The other part deals with the reduced perception of color breakup on the novel spectrum‐sequential display compared to conventional color‐sequential displays.     

Relationship between mode‐boundary from surface color to fluorescent appearance and preferred gamut on wide‐gamut displays

Abstract— Color‐gamut design is a major concern in wide‐gamut displays. To determine a preferred gamut for displaying object color in natural scenes on a wide‐gamut display, subjective evaluations were conducted to investigate the preferred color and acceptable limit. Then, simple synthesized images were used to determine the mode boundary between surface color and fluorescent color appearance. It was found that (1) observers perceived the colors with high saturation and high lightness as fluorescent colors and (2) the fluorescent appearance decreased preference. The color‐mode index (CMI) was defined as an evaluation index of the color‐appearance mode so that the boundary between surface color and fluorescent color appearance was defined as CMI 100. Additionally, it was found that the CMI 100 loci could be interpreted as an optimal color loci. Then, it was clarified that the mode boundary and the preferred gamut were closely related and that the acceptable limit for L* was 1.1 times L* for CMI 100.