Abstract: This paper presents a novel approach for edge-based shape retrieval for improving the conventional algorithm of image retrieval based on edge orientation histogram. It extracts the image edge by the algorithm of the wavelet transform based on maximum model. It gets the coherence edge direction histogram which depends on the edge point’s directions and correlation between neighboring edges. It counts the similarity by circular shift, and then applies it to image retrieval. Experimental results show that the algorithm has the robustness against illumination changes, position translation, scale and rotated image and a superior image retrieving performance.

Key words: image retrieval, edge direction histogram, coherence edge orientation histogram, shape representation

摘要： 针对传统的边缘方向直方图在图像检索应用中的不足，提出了一种基于聚合边缘方向直方图的图像检索算法。采用小波模极大值边缘检测算法提取图像边缘；根据图像边缘点的方向信息及其之间的空间分布特征得到聚合边缘方向直方图；通过循环移位计算直方图间的最小距离作为相似度，实现了图像检索。该方法具有较好的平移、旋转、尺度不变性。实验结果证明了该算法的有效性和可行性，并取得了较好的检索效果。

关键词: 图像检索, 边缘方向直方图, 聚合边缘方向直方图, 形状描述