The RAMSES platform for a triple interactive screen: application to the automatic generation of interactive television.

Summary With the digital revolution, the use of video has strongly evolved during the last decades, from cinema to television, then to the web, from fictional storytelling to documentary and from editorialization to user creation. Media are the vectors for exchanging information, knowledge, personal "reports", emotions. The automatic enrichment of multimedia documents is still a subject of research since the advent of the media. In this context, we first propose a modeling of the different concepts and actors implemented to automatically analyze multimedia documents in order to dynamically deploy interactive services in relation with the media content. We define the concepts of parser, interactive service, description of a multimedia document and finally the functions necessary to make them interact. The resulting analysis model stands out from the literature by proposing a modular, open and scalable architecture. We then present the implementation of these concepts in a demonstration prototype. This prototype allows us to highlight the advanced contributions in the description of the models. An implementation and recommendations are detailed for each model. In order to show the results of the implementation of the solutions proposed on the platform such as MPEG-7 standards for the description, MPEG-4 BIFS for the interactive scenes or OSGI for the general architecture, we present different examples of interactive services integrated in the platform. These examples allow us to verify the ability to adapt to the needs of one or more interactive services.