TUT Audio Event Detection System 2007

This paper describes a system used in acoustic event detection task of the CLEAR2007 evaluation. The objective of the task is to detect acoustic events (door slam, steps, paper wrapping etc.) using acoustic data from a multiple microphone set up in the meeting room environment. A system based on hidden Markov models and multichannel audio data was implemented. Mel-Frequency Cepstral Coe cients are used to represent the power spectrum of the acoustic signal. Fully-connected three-state hidden Markov models are trained for 12 acoustic events and one-state models are trained for speech, silence, and unknown events. The system performed adequately compared to other participant in CLEAR2007 evaluation.