Speaker diarization is originally defined as the task of de- termining “who spoke when” given an audio track and no other prior knowledge of any kind. The following article shows a multi-modal approach where we improve a state- of-the-art speaker diarization system by combining standard acoustic features (MFCCs) with compressed domain video features. The approach is evaluated on over 4.5 hours of the publicly available AMI meetings dataset which contains challenges such as people standing up and walking out of the room. We show a consistent improvement of about 34 % rela- tive in speaker error rate (21 % DER) compared to a state-of- the-art audio-only baseline.