Research on Multimedia Content Production for telelecture between China and Japan

点击翻译按钮获取中文摘要

This research covered a model system for telelecture application between China and Japan in the first year. Issues on a) system environment, b) media input flow, c) feature extraction and description of multimedia were investigated. And in the second year, a telecturing system between Beijing University of technology and Waseda University was designed and implemented based on the first year's result. Using the implemented web-based multimedia lecture system, 2 major topics, a) accessibility enhancement to the requested content and material and b) improvement of data transmission, were studied in the following 4 subjects.1. Rotation and shift free Content-based image retrieval method: As a new feature "disorder matrix" was defined as an independent parameter from rotation and spatial shift of the content. Using disorder matrix and global color histogram, fast selection of matching candidates has become possible. From the simulation experiments, the proposed algorithm's efficiency in cal … More culation time and robustness was confirmed.2. Progressive image coding scheme based on DCT coefficients: We have developed a new progressive image coding scheme, by combining DCT and SPECK algorithm. The new scheme's most advantageous point is its low computational cost and will be applicable for content browsing on the low performance terminals.3. Implementation of eXtensible Lecture On Demand based on LML (Lecture Markup Language): LML is designed to enable multimedia synchronization as SMIL and based on XML. In this research, Lecture on Demand system was implemented using LML and carried out subjective evaluation on XML. In this research, Lecture on Demand system was implemented using LML and carried out subjective evaluation test of the system.4. Intelligent multimedia content generation systems: In this study, 2 systems were designed and implemented as an application of image processing and feature description technologies, which were a) higher resolutional ROI (Region of Interest) image generation system by video mosaicing and b) high efficient web-based video authoring system using camera working parameters. These two systems provide more functions than the user's tool's capabilities by the help of image processing technologies and will be useful especially in various telelecturing occasions. Less