Semi-supervised On-line Speaker Diarization for Meeting Data with Incremental Maximum A-posteriori Adaptation | Publicación