웹 기반 멀티미디어 컨텐츠 개발에 있어 인덱싱 기법

Indexing Method in Web Based Multimedia Content Development

Every mans movements can be useful data. In broad definition, we can call data as content. In order to utilize people‘s experience and know-how, and eventually share (through training and education) them we have to record human activities. Currently we record these data as video material because its educational result is better. In addition, the video method is becoming more cost effective and easier to produce. To share and transfer video records, requires a more sophisticated and complicated recording process. In other words, to provide on-line video materials requires very precise modification procedures and techniques which will show only selected parts of the video. With video tape the user must fast forward and rewind to find the desired part. In order to use digitalized video records, documents, or other types of supplementary records we must find a new methodology. An experiment is designed to discover the right methodology to utilize contents effectively (from development to usage), and prove its effectiveness. The focus is on the meaningful content division and content-based indexing method on the web. If we have a well established information management system, it will become a strong point and will result in multiple business opportunities. Among the various kinds of multimedia data, video and audio contents are the two kinds that have a continuing flow of data with starting and ending times. This continuation of audio and video can be divided into certain levels of important units. For example, the user can designate the smallest unit of the content and store the starting and ending information. There are key words inside of flow of video and audio contents, which will tell users about the content and can be used as a clue for users to activate a revisit. The user wants to search only the needed part as fast and easy as possible not bothered by the time and space simultaneously. In order to satisfy the users desire, we have to modify the contents (modeling & indexing) in advance and put it in the database with a special technique. This experiment shows what kind of methodology we need and actually provides proof based on that theory.