IHLS颜色空间的灰度化全局非线性三向映射

Decolorize Nonlinear Tri-mapping in IHLS Color Space

为了减少彩色图像转化为灰度图像过程中颜色信息的损失,基于人眼感知明度的视觉模型,提出一种全局灰度化方法.首先使用彩色图像的IHLS颜色空间,建立了关于亮度、色度和饱和度3个因子的全局非线性映射模型;然后将该映射模型中全局参数的计算转化为一个能量函数的优化,其中第一个能量项度量了彩色图像和灰度图像梯度的差异,第二个能量项定义了图像中给定一个像素与其随机采样点集之间两两彩色距离与灰度距离的差异;最小化上述2个能量项可以通过3个因子的迭代算法进行求解.大量的实验结果表明,该方法可更好地反映人眼对彩色图像的感知,使得灰度转化结果更加贴近颜色的感知明度.

In order to reduce the loss of color information, we present a global nonlinear tri-mapping approach for converting color images into grayscale ones, which is based on the perceived intensity experiments. This conversion adopts the IHLS color space and builds one global nonlinear mapping with three attributes, i. e. , luminance, hue, and saturation. The global parameters are obtained by minimizing an energy function, where the first term measures the difference of color gradient and gray gradient, and the second one defines the difference of the color and gray distance between one pixel and its random sampling. This optimization method achieves the minimum by the iteration separately with three attributes. Experimental results show that our method produces pleasing results, which are close to the perceived lightness of the original color images, for a variety of color images.