基于内容的标准烟叶图像数据库检索

基于内容的检索是当前多媒体数据库发展的一个重要研究领域。在图像数据库中,基于内容的图像检索技术一般采用颜色直方图为特征。该文提出了把图像的形状特征、颜色特征和纹理特征结合起来的方法进行基于内容的图像检索。通过对人的视觉检索过程的研究,提出把数据库中的图像依次分别按形状特征、颜色特征和纹理特征分级聚类的方法,既符合人的视觉特点又大大提高了检索效率。

Content Based Image Retrieval of Standard Tobacco Leaf Database

This paper deals with efficient retrieval of images from standard tobacco leaf database.Content based image retrieval is an important study field in multimedia database.In an image database,image retrieval is usually based on color histograms ,but this method can not preserve the spatial information in the image.In this paper,the technique combines features of shape,color and texture in images,this method can either preserve statistical features or spatial information in the image.The paper puts forward a classification clustering method that clustering images according to shape,color and texture feature in turn.This method can greatly improve retrieval efficiency.People use this method to retrieval images in the standard tobacco leaf database,and the experiments show that this approach can get good result.