Level 3 of the display specification of Teletext and Videotex includes the provision of dynamically redefinable character sets which can be used to construct new characters or arbitrary graphic images. The authors describe how character recognition and clustering algorithms can be used to map from a pixel-based image to the level-three character-oriented Teletext/Videotex display format. The mapping typically requires a compromise to be made between image coding efficiency, image accuracy, and aesthetic factors and would require an interactive system where the image originator retains aesthetic control while varying the character recognition and clustering parameters. An overview of image analysis for level-three Teletext is given. Pattern matching using standard character sets is described.
展开▼