融合颜色与形状特征的图像检索方法  

Image Retrieval Based on Fusion of Color and Shape Features

作　　者：卢成武[1] 种兰祥[1] 

机构地区：[1]西北大学电子科学系

出　　处：《微电子学与计算机》2006年第2期160-162,共3页Microelectronics & Computer

摘　　要：基于颜色或颜色-空间信息的图像检索方法,由于没有考虑图像中所含目标对象的形状特征,检索效果往往不够理想,针对这一不足。文章提出并设计了颜色-梯度方向角二维直方图,将图像的颜色特征与形状特征融合起来进行图像检索。试验结果表明,该方法的检索精度与效率都有明显的提高。The methods that based on color or color-space features for image retrieval are simple and efficiently, but no consider the shape feature of the object in the image. So its retrieval result always not ideally. As to this practical problem, we propose a new image retrieval method which base on color-Directional Gradient Angular 2D Histogram. The experiment indicate that this method is great effective in image＇s retrieval and improves the image retrieval precision.

关 键 词：HSV 形状特征 梯度方向角 直方图 图像检索 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]