Speaker Diarization Using Direction of Arrival Estimate and Acoustic Feature Information: The I2R-NTU Submission for the NIST RT 2007 Evaluation

This paper describes the IR/NTU system submitted for the NIST Rich Transcription 2007 (RT-07) Meeting Recognition evaluation Multiple Distant Microphone (MDM) task. In our system, speaker turn detection and clustering is done using Direction of Arrival (DOA) information. Purification of the resultant speaker clusters is then done by performing GMM modeling on acoustic features. As a final step, nonspeech & silence removal is done. Our system achieved a competitive overall DER of 15.32% for the NIST Rich Transcription 2007 evaluation task.