摘要
基于DICOM标准的PACS系统是世界医学信息技术发展的潮流.文章具体分析元数据检索、文本检索和基于内容的图像检索的优缺点,认为有必要在牙科PACS中综合使用三种检索方式以提高信息检索效率和准确度.在基于内容的图像检索中,为了提高在庞大的特征库中进行相似性检索的效率,需要研究快速索引结构.文章通过分析R树、四叉树和聚类索引等几种快速索引结构,初步设计出适合牙科PACS系统的索引机制.
PACS based on DICOM is the trend of world's medical information technology development. This paper analyzes in detail three kinds of image retrieval (metadata, textual description and content-based image retrieval) as well as their advantages and disadvantages. It can be concluded that combining the three kinds of image retrieval helps formulate more precise queries in dental PACS. By doing so, we hope to achieve better results. When using CBIR, it's necessary to study fast index structure to improve similarity retrieval efficiency in large feature database. After analyzing the principles, advantages and disadvantages of several kinds of fast index structure, including R-tree, Quad-tree, and clustering indexing, the paper designs an index mechanism suitable for the dental PACS.
出处
《西南民族大学学报(自然科学版)》
CAS
2007年第4期976-979,共4页
Journal of Southwest Minzu University(Natural Science Edition)
基金
成都信息工程学院2006年度自然科学与技术发展基金项目(CSRF200604).