Abstract: A color image segmentation algorithm based on S-CIELAB color space is presented in this paper. The approach transforms the RGB image into S-CIELAB color space based on color transfer characteristics of human visual system, the image is segmented using mean shift algorithm. Experimental results show that the approach simulates the eye blurring mechanism by filtering the origin image, which helps to produce result similar to human perception. The color image can be segmented effectively by using the present method even if it is affected seriously by Gaussian noise.

Key words: color image segmentation, pattern-color separable model, color space, mean shift algorithm

摘要： 提出一种基于S-CIELAB颜色空间的彩色图像分割算法。在人类视觉彩色传递模型的基础上，将原始的RGB图像转换到S-CIELAB空间，运用均值漂移算法对图像进行分割。实验结果表明，该算法能模拟人类视觉模糊特性，得到与人类视觉非常接近的分割结果。对于被高斯噪声严重污染的彩色图像，该算法也能有效地进行分割。

关键词: 彩色图像分割, 模式-彩色分离模型, 颜色空间, 均值漂移算法

CLC Number: 

• TP391