摘要
提出了基于多尺度的边缘链码方向的图像检索算法。首先采用不同的尺度对图像进行轮廓描述,获得多尺度的图像轮廓图;再对不同尺度的轮廓图进行链码方向统计,计算不同链码方向所占的比值,计算均向量;最后利用向量的欧式距离度量相似性。实验结果表明,该算法具有查准率高、检索性能良好等优点。
Image retrieval algorithm based on multi-scale marginal chain code orientation was proposed.First,the multi-scale image contour map was obtained by outlining the whole image with different scales.Second,the proportion of each orientation was calculated,that was obtained by counting the map of different chain code orientations.And the mean vector was computed.Finally,the vectorial Euclidean distance was used to measure the similarity.Experimental results showed that the algorithm has merits of high precision and retrieval performance.
引文
[1]COSTA L F,CESAR R M.Shape Analysis and Classification:Theory and Practice[M].State of Florida:CRC Press Inc,2001:1-685.
[2]JIANG X D,YAU W Y.Fingerprint Minutiae Matching Based on the Local and Global Structures[C]//15th International Conference on Pattern Recognition,Quebec City,Canada,August 11-15,2000,New York:IEEE,2000,2:1038-1041.
[3]SCHIETSE J,EAKINS J P,VELTKAMP R C.Practice and Challenges in Trademark Image Retrieval[C]//Proceedings of the 6th ACM International Conference on Image and Video Retrieval.Amsterdam,Netherlands,July 9-11,2007,New York:ACM,2007:518-524.
[4]刘威,孔英蕾,肖宏涛.基于形状特征的图像检索技术[J].兰州理工大学学报,2011,37(1):100-105.
[5]HERMILO S C,Ernesto B,Ramon M R D.Efficiency of Chain Codes to Represent Binary Objects[J].Pattern Recognition,2007,40(6):1660-1674.
[6]余博,郭雷,赵天云,等.Freeman链码描述的曲线匹配方法[J].计算机工程与应用,2012,48(4):5-8.