Support System for Telecommunication Using Keywords

We developed a prototype system to support telecommunications that uses keywords selected by the presenter in videoconference. The presenter holds to show keyword cards, and the listeners at remote sites can see additional information with the keywords. Although keyword captions are considered effective in video learning environments for learning foreign languages, we think they should also be available for interactive communications. Our prototype system recognizes letters and characters in a video image, and provides us with additional functions, such as language translation, 3D models, and audio reproductions. The visual data are overlaid onto the real scene with a multilayered display, using a Web browser as a presentation tool to enable us to easily author/edit multimedia data. Optical character reader (OCR) middleware was implemented into the recognition function for Japanese, and an open-source library, ARToolkit was used to detect the character markers that controlled the functions of the information presentation. The support system is designed to efficiently exploit audiovisual signals of existing videoconferencing systems