摘要
提出了一种新的基于边缘颜色分布的图像检索算法。该算法将不同类型边缘附近的颜色分布作为刻画图像内容的主要特征,设计了一种紧凑的2D边缘颜色直方图来对图像的边缘颜色分布特征进行描述,既利用了局部颜色特征,又考虑了不同类型边缘的空间分布信息,克服了传统颜色直方图不能反映空间信息的缺陷。实验结果表明,该算法与其他同类方法相比,有效地提高了检索性能。
A new image retrieval algorithm based on edge color distribution was proposed. The content information of an image was characterized by color distribution of different edge types described by two-dimension edge color histogram. The algorithm takes not only the local color features into account but also the space distribution information of edge types and overcomes the shortcoming of color histogram which loses the location information. Experimental results show that this method effectively improves the image retrieval performance compared with other retrieval methods.
出处
《计算机应用》
CSCD
北大核心
2007年第7期1651-1653,共3页
journal of Computer Applications
基金
国家863计划项目(2006AA01Z127)
国家自然科学基金资助项目(60572152)
陕西省自然科学基金资助项目(2005F26)
关键词
基于内容的图像检索
特征提取
边缘颜色分布
Content-Based Image Retrieval (CBIR)
feature exaction
Edge Color Distribution (ECD)