摘要
图像特征的提取和使用在基于内容的图像检索中至关重要。研究了在基于内容的图像检索系统中整合颜色,纹理,形状的提取方法。将图像按照一定的规则进行分块,对各个分块分别进行各种特征向量的提取。颜色特征的提取是基于YUV颜色空间的颜色直方图,纹理特征的提取采用Gabor滤波器,形状特征的提取是基于Zernike矩的计算。实验结果表明,综合图像的颜色、形状和纹理特征提高了图像检索的准确性。
Image feature extraction and the use of the features as query terms are crucial problems in content-based image retrieval （CBIR） systems. The main focus is on integrated color, texture and shape extraction methods for CBIR. The image is devided into blocks according to regular, then feature is extracted vector in every block. Color features extraction is based on color histograms in YUV space, texture feature based on Gabor filters and shape features based on Zernike moments are calculated. The experimental result indicates that the arithmetic of the extraction based on the feature of color, texture and feature improve image retrieval accuracy.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第21期4904-4906,共3页
Computer Engineering and Design
作者简介
孙文波（1982-），男，山东威海人，硕士，研究方向为人工智能与模式识别；
吴锡生（1968-），男，江苏无锡人，博士，教授，研究方向为人工智能与模式识别。E-mail：sunwenbobo1357@126．com
