| 图像数据库基于内容检索的索引方法研究 |
| |
| 引用本文： | 林坤辉,徐焕,周昌乐.图像数据库基于内容检索的索引方法研究[J].厦门大学学报(自然科学版),2006,45(4):488-491. |
| |
| 作者姓名： | 林坤辉 徐焕 周昌乐 |
| |
| 作者单位： | 1. 厦门大学软件学院
2. 厦门大学信息科学与技术学院,福建,厦门,361005 |
| |
| 基金项目： | 厦门大学“985”二期信息创新平台项目资助 |
| |
| 摘 要： | 为了提高图像数据库的检索效率，必须提高高维索引的效率．通过对SR-tree和x-tree的结构和性能分析，引入X-tree中超级节点的思想，改进了分裂算法，设计了一种新的高维索引结构ESR-tree（Extended SR-tree）．ESR-tree采用超矩形和超球形相结合的包络方法，在节点结构中引入超级节点．通过改进插入和分裂算法，有效降低了重叠率，避免了不必要的分裂，更好地维持树的平衡．同时有效降低了CPU时间和I／O次数，提高了检索效率．实验表明，随着数据量和维数的增多，ESR-tree的性能明显优于SR-tree和X-tree．
|
| 关 键 词： | 基于内容检索 高维索引 超级节点 |
| 文章编号： | 0438-0479(2006)04-0488-04 |
| 收稿时间： | 11 30 2005 12:00AM |
| 修稿时间： | 2005年11月30 |
Image Database Content-based Retrieval Indexing Analyses |
| |
| LIN Kun-hui,XU Huan,ZHOU Chang-le.Image Database Content-based Retrieval Indexing Analyses[J].Journal of Xiamen University(Natural Science),2006,45(4):488-491. |
| |
| Authors: | LIN Kun-hui XU Huan ZHOU Chang-le |
| |
| Institution: | 1. School of Software,Xiamen University, 2. School of Information Science and Tchnology,Xiamen University, Xiamen 361005 ,China |
| |
| Abstract: | The enhancement of high-dimensional indexing technique is necessary to improve the performance of image database retrieval.In this paper,the structure and performance of SR-tree and X-tree was analyzed,and split algorithm was improved by introducing the super-node idea in X-tree.A new high-dimensional indexing structure ESR-tree(Extended SR-tree) which used super-sphere and super-rectangle as data package was designed.It reduced overlap effectively,avoided unnecessary split and maintained the balances of the indexing tree.It also reduces the CPU time and I/O number and improves the retrieval efficiency.Experiments show that the performance of ESR-tree is better than SR-tree and X-tree with the increasing data number and dimensions. |
| |
| Keywords: | content-based retrieval high-dimensional indexing super-node |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
