一种基于图像特征的图像分类方法  

Algorithm of Image Classification Based on Image Feature

作　　者：杨怿菲[1] 

机构地区：[1]西安邮电学院,陕西西安710121

出　　处：《现代电子技术》2009年第14期81-82,86,共3页Modern Electronics Technique

基　　金：陕西省自然科学基金资助项目(2006F17)

摘　　要：图像分类是色域匹配的关键环节,不同类型的图像采用不同的匹配方法。针对如何有效分类图像,设计了一种基于图像特征的图像分类算法。首先建立图像颜色的三个通道特征统计模型和基于空间灰度级的纹理统计、边缘特征的统计模型,然后根据模型计算出图像的三类特征值,利用特征统计评判和神经网络技术分析计算数据,最后得出图像类型。实验结果表明,算法有较高的分类精度。Image classification is a key technology for gamut mapping. Different method is used or different image type. For how to effectively classified image, an algorithm of classification based on image feature is designed. Three - channel characteristics of a statistical model of image color, gray - scale space - based texture statistics model and image edge feature model are given. The calculation results of models are organized to judge the type of image by statistics and neural network technology, then image types are determined. The algorithm has higher accuracy by experimental data.

关 键 词：图像特征 图像分类 颜色 纹理 边缘特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]