The ESTER Evaluation Campaign for the Rich Transcription of French Broadcast News

This paper gives an overview of the ESTER evaluation campaign. The aim of this campaign is to evaluate automatic broadcast news transcription systems for the French language. The evaluation tasks are divided into three main categories: orthographic transcription, event detection and tracking (e.g. speech vs. music, speaker tracking), and information extraction (e.g. named entity detection, topic tracking). Each category is evaluated separately. This paper gives details about the tasks to be performed and the corpus, with particular emphasis on the manually transcribed reference transcription. 1