一种新的颜色空间量化方案  

A Novel Color Space Quantization Scheme

作　　者：周咏梅[1] 徐德智[1] 阳爱民[2] 朱艳辉[2] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083 [2]湖南工业大学计算机科学与技术系,湖南株洲412008

出　　处：《株洲工学院学报》2006年第4期39-42,共4页Journal of Zhuzhou Institute of Technology

基　　金：湖南省自然科学基金资助项目(05JJ40101);湖南省教育厅基金资助项目(03C597)

摘　　要：详细介绍样本颜色的采集方法、采集软件的设计以及数据的处理方法;提出了基于BP神经网络的样本颜色处理方法,介绍其在图像特征的抽取、区域分割的应用。In this paper, a novel color space quantization scheme with human perception for the different color during the quantization process is proposed. The paper introduces the method of collecting color samples and the design of collection software and presents a method of data processing based on BP neural network for color samples. This quantization method is used for color features extraction and image segmentation based on region. The method of considering human perception into image retrieval system accords with human vision habits and reduces the gap between low-feature and high-level semantic concepts in a new perspective.

关 键 词：颜色空间量化 颜色样本采集 BP神经网络 图像区域分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]