摘要
随着音频检索需求的提高,传统数据库越来越不能满足音频检索的需求,基于内容的音频检索的研究变得越来越重要。因此对音频特征的直方图表示及音频索引二叉树进行了调查,在此基础上详细描述了基于音频索引二叉树的检索算法(包括统一格式算法、非统一格式算法、部分匹配算法、虚拟节点算法),分析这些算法的时间复杂度,为下一步研究做准备。
Content-based indexing of audio (and multimedia) data has become more important since conventional databases cannot provide the necessary efficiency and performance. The histogram-based feature and the construction of the binary tree of a histogram-based feature are introduced. The algorithms based on the binary tree of a histogram-based feature are introduced, including search-by- restricted-format (SBRF) algorithm, search-by-unrestricted-formatted (SBUF) algorithm, the partial-matching (PM) algorithm and the virtual-node (VN) algorithm. And the time complexity of the algorithms is analyzed.
出处
《计算机工程与设计》
CSCD
北大核心
2006年第16期3003-3006,共4页
Computer Engineering and Design