Finding elements of interest within a recorded meeting is time-consuming. Using the AMI meeting archive browser, JFerret, we support this process by displaying many types of data. Data types include synchronized media, slides and transcripts. We can also display processing results, such as a “synthesized” agenda, speaker segmentations, diarization and argumentation. Users interact with these visualizations to find elements and control synchronized playback of the recorded meeting. The extension of this is to be able to browse a corpus of meetings and to search among many meetings to find the most relevant pieces to answer a particular question. [demonstration]

Click on the image to see full size