Segmentation of TV shows into scenes using speaker diarization and speech recognition | IEEE Conference Publication | IEEE Xplore