基于在线学习的彩色图像区域增长法

Color Images Region Growing Method Based on Online Learning

论文阐述了一种基于在线学习算法的彩色图像区域增长法,用于解决基于内容的图像检索系统。该算法采用贝叶斯估计法的变分描述每一个增长区域,由此得出的图像生成过程应用简单,初始参数具有鲁棒性,且具有线性复杂性。图像处理结果表明了该方法用于基于内容的图像检索系统的可行性。

This paper presents a color images region growing method based on online learning algorithm,which is used for content-based image retrieval systems.The algorithm adopts a variation of Bayesian estimation procedure to characterize each grown region.The resulting image growing procedure is simple to implement,robust to initial parameters and has a linear complexity.Results on several images show the feasibility of the proposed method,which is used in the content- based image retrieval systems.