Texmix: an Automatically Generated News Navigation Portal Illustration 1: Overview of the Navigation Interface

The Texmix demonstration presents an original interface to navigate a collection of broadcast news shows, exploiting speech transcription, natural language processing and image retrieval techniques. Navigation is performed through keywords search or through time or through maps, with links automatically created either between reports to follow the story or to the Web to know more about a story, a person or a fact. Image search technology is also integrated to find portions of the collection with similar images. We also present two original features to dynamically access videos, namely dynamic summary and geotagging. 1. INTRODUCTION The gradual migration of television from broadcast diffusion to Internet diffusion offers tremendous possibilities for the generation of rich navigable contents. However, it also raises numerous scientific issues regarding delinearization of TV streams and content enrichment. In this demonstration, we illustrate how speech in TV news shows can be exploited for delinearization of the TV stream. In this context, delinearization consists in automatically converting a collection of video files extracted from the TV stream into a navigable portal on the Internet where users can directly access specific stories or follow their evolution in an intuitive manner. The TEXMIX demonstration illustratess the result of the delinearization processand enables user to experience various navigation modes within a collection of news shows. An illustration of the interface is provided Figure 1. The demonstration can be accessed online at http://texmix.irisa.fr. Structuring a collection of news shows requires some level of semantic understanding of the content in order to segment showsinto their successive stories and to create links between stories inthe collection, or between stories and related resources on the Web. Spoken material embedded in videos, accessible by means of automatic speech recognition, is a key feature to semantic description of video contents. At IRISA/INRIA Rennes, we have developped multimedia content analysis technology combining automatic speech recognition, natural language processing and information retrieval to automatically create a fully navigable news portal from a collection of video files [1]. The purpose of this demonstration, aside from exploiting technologies developed in our research group, is to provide a new experience from watching videos on the Internet. Exploiting HMTL5 features, dynamic content is provided as the video plays, and allow the watcher to have additional information and to interact during his watching, being no longer passive. We also provide new search services based on the technologies developed. Permission to make digital or hard …