Fumio Hasegawa

A video metadata application and its verification test using NIPEG-7 description and watermarked content ID

Because of the rapid progress of broadband networks and the rise of digital broadcasting, there is a great demand for creation of an environment for digital content production and establishment of system for smooth content distribution. Yamagata Video Archive Research Center (YRC) has been involved in the development of technology for a large-scale video archive system and in research activities, through various verification tests, for applicable operation of the system. We adopt a MPEG-7 description scheme for constructing retrieval keywords in order to deal with video distribution and to achieve easy content retrieval for users outside of YRC In addition, we use a watermarked unique code (cid) system propounded by the Content ID Forum (CIDF) for managing the content identification number. A part of content ID is watermarked on video itself and a time code, which is extracted from an original tape, is watermarked for editing purposes to keep MPEG picture quality in an edited video stream. We describe the research through tests from content retrieval to content distribution.

Petabyte-class video archive system with large-scale automatic tape feed robot

In order to supply a volume of digital video content in the digitalization of broadcasting, it is necessary to realize a method of storing a large number of videos and of high-speed content retrieval. This paper describes a petabyte (1015 power) large-capacity video archive system and its retrieval method. The archive system employs an automatic magnetic tape feed robot that can accommodate 7200 tapes (15,000 hours) in its center, and an HDD server and a DVD server respectively for each purpose, thus making the efficiency of storage and distribution better and realizing retrieval, video delivery, and remote editing from remote sites. Further, a method of detection of scene change points in a video sequence, a method of browsing video production based on the detected scene change points, and a method of visual retrieval by simultaneous playback representation of multiple videos on a screen are also included. In this paper, a large-scale archive system that has an automatic tape feed robot at its center and a high-speed and efficient method of retrieval of the target video using the real-time playback representation of multiple videos are presented.

Content Creation Utilizing a Large-scale Archive with Content ID and Original Time Code

Stored video data in large-scale archives will be used primarily for content creation and, second, for content distribution. In both cases the protection of intellectual property is an indispensable issue. A good solution of this issue may be to adopt the Content ID (identification) being promoted by the Content ID Forum. This is a unique identifier to realize some benefits of materials (or contents) distribution. A part of the Content ID will be watermarked on the video itself and will include a time code from an original tape (or cassette). Preview of materials in a data-networked environment is another issue. MPEG-4 was chosen as the compression technology because of its lower-band transmission capability; the Content ID could be easily carried with its stream. A remote editing terminal outputs an edit decision list (EDL) to an editing studio for high-quality mastering. In this process, the EDL is created with the original time code. On the other hand, the local time code might be used in the editing studio, in which case, some time code conversion is needed to execute the EDL.