|
|
论文题目 |
对等网络中基于语义的音乐信息融合与检索系统研究 |
论文题目(英文) |
|
作者 |
王珺 |
发表年度 |
2008 |
卷 |
29 |
期 |
9 |
页码 |
63-69 |
期刊名称 |
微计算机应用 |
摘要 |
随着对等网络中数字音乐内容的丰富,音乐信息检索日益成为对等网络音乐共享系统中的一个关键问题。针对对等网络环境,本文提出了一种基于语义的音乐信息融合与检索系统。文中首先定义了一种可拓展的音乐本体;然后提出了将自动提取的特征信息、用户标注的信息以及网络获取的信息等多信息源进行融合的方法,并且介绍了在CLAM的Annotator模块基 础上开发的融合标注工具;继而提出了一种基于RDFPeers架构的静态及动态信息的存储和检索方法。本系统使对等网络中的音乐信息检索从传统的基于文件名、艺术家名等关键字的检索拓展到基于语义的音乐信息检索,同时对该系统的示例模型的评估也验证了本文提出的多信息源融合的方法能够优化系统的准确率和召回率。 关键词:对等网络音乐信息检索低层描述等 |
摘要_英文 |
As the digital music collections in Peer—to—Peer(P2P)network expand larger and larger,music information retrieval(MIR)is becoming a crucial problem in the P2P music—sharing systems.We propose a semantics—based music information aggrega-tion and retrieval system especially for P2P network.First,anextendable music ontology is defined;then we bring the pmposal of aggregating the multi—information which contain auto—extracted features,user—annotated descriptors and Web metadata,etc.,and in·traduce the relatedextractor—aggregator tool developed based OH CLAM Annotator module;furthermore,a method based on the RDF-Peers framework is introducedto repesit and query both the static and the dy,mmic information.The proposed system extends the MIR iIIP2P network from the traditional title/name keyword retrieval to semantics—based information retrievaL The evaluation of the demonstration model also shows that the proposed multi—information aggregation mentod optimizes the retrieval accuracy and recall Keywords:Semantics。P2P MIR,Low—level descriptors |
|
|
|