FIT Božetěchova 2, seminární místnost UPGM L220 23.6.2008

On this seminar talk, a recently proposed method of joint pitch (Fo) and direction of arrival (DoA) extraction for speaker localization will be introduced. Beside the basic idea of the multidimensional Position-Pitch (PoPi) decomposition, its application in surveillance and hands-free communication systems will be discussed.
The proposed method will be demonstrated on concurrent speaker and moving speaker scenarios on real-world recordings and compared to state-of-art DoA estimation methods. Comparison results will show the power of the method in determining the accurate pitch and position estimates for multi-source scenarios. We will see that tje Position-Pitch decomposition (PoPi) gives an intuitive representation of all active speakers in terms of their respective position and pitch estimates.