Natural language based approach to video indexing

概要

論文の詳細を見る
This paper describes methods in which natural language is used to describe video contents, knowledge of which is needed for intelligent video manipulation. The content encoded by natural language is extracted by a language analyzer in the form of subject-centered dependency structures and is combined in an incremental way into a single structure called a multi-path index tree. Descriptors and their dependency relations are extracted from the index tree in order to provide retrieval flexibility, content-based video index is eventually represented in a two-dimensional structure in where descriptors and relations are mapped onto a component axis and temporal references are mapped onto a time axis. The proposed index structure l) is suitable for queries by keywords or by natural language, 2) can be generated by an automated procedure, and 3) has a compact and flexible structure that is easily expandable. Integration with vision processing will enable to also handle nonsymbolic data such as the color, shape, and directions of visual objects.
社団法人電子情報通信学会の論文
1995-10-16