基于色彩纹理特征的图像分类及应用

A study on classification and application of images based on color texture feature

目的研究彩色数字图像的计算机分类识别方法并应用于古瓷片的自动分类。方法提出了一种色彩纹理特征的提取模型,采用该模型,利用IGabor滤波器提取数字图像的色彩纹理特征,并构造支持向量分类机(SVM)分类器组。结果实现了高准确率多类别图像的自动分类识别,并成功应用于古瓷片的自动分类。结论色彩纹理特征提取方法将颜色与纹理进行融合,增强了数字图像之间的特征区分能力。

Aim To stady automatic classification methods of digital image and apply the methods to ancient shards.Methods Give a method for extract image Color-Texture features,use IGabor filter-banks to extract Texture-Color features from images,and then construct a group of support vctor machines（SVM） classifiers.Results Automatically and accurately the color digital images have been classified and successful experiments been applied to ancient shards image.Conclusion Integrations of the texture features and color information can further improve distinguishing ability of each category images.