首页>
外国专利>
Automatic indexing of digital image archives for content-based, context-sensitive search
Automatic indexing of digital image archives for content-based, context-sensitive search
展开▼
机译:自动索引数字图像档案库,用于基于内容的上下文相关搜索
展开▼
页面导航
摘要
著录项
相似文献
摘要
The present invention relates to a method for automatically indexing multimedia data archives (304c) and categorizing the files (102a-7) held therein, and a client / server architecture of an image retrieval system (300) for content-based searching for relevant files (102i) of a particular one Formats and a specific file structure. DOLLAR A According to the invention, the files (102a-n) stored and parsed in a file archive (304c) managed by the document management system (304) are subjected to a feature extraction algorithm (S0a). The extracted features are then used to generate a binary-coded inverted index (116) consisting of elements of at least two attributes, in which context information contained in these files (102a-n) is provided for content-based image retrieval. If new files (102o-z) or files (102a-n) with extended or modified content are stored in the data archive (304), according to the invention a parsing algorithm (S0b) and an algorithm (S0c) for automatic extraction are created for each storage operation features of these files (102a-z) are event-driven to augment the inverted index (116) with individual attributes or to update certain elements of existing index attributes.
展开▼