160 likes | 294 Views
Source Number Estimation and Clustering for Undetermined Blind Source Separation. Benedikt Loesch and Bin Yang University of Stuttgart Chair of System Theory and Signal Processing International Workshop on Acoustic Echo and Noise Control, 2008. Presenter Chia-Cheng Chen. Outline.
E N D
Source Number Estimation and Clustering for Undetermined BlindSource Separation Benedikt Loesch and Bin Yang University of Stuttgart Chair of System Theory and Signal Processing International Workshop on Acoustic Echo and Noise Control, 2008 Presenter Chia-Cheng Chen
Outline • Introduction • Observation Vector Clustering • Source Number Estimation • Experimental results • Conclusion
Introduction • The task of blind source separation is to separate M (possibly) convolutive mixtures xm[i],m = 1, . . . ,M into N different source signals. • Present an algorithm call NOSET (Number of Source Estimation Technique)
Observation Vector Clustering(1/3) • Short Time Fourier transform (STFT) • Three steps • Normalization • Clustering • Reconstruction
Observation Vector Clustering(2/3) • Normalization • The normalization is performed with respect to a reference sensor J [4] • Unit-norm normalization
Observation Vector Clustering(3/3) • Clustering • K-means • Reconstruction [4]
Source Number Estimation(1/6) • The phase difference among different sensors is large enough. Inthe low-frequency region, this is not the case and the phase estimate is rather noisy. • Only one source is dominant at a TF point [k, l].
Source Number Estimation(2/6) • Selection of One-Source TF Points • Power of source n • Selection of reliable TF points
Source Number Estimation(3/6) • DOA Estimation • time delay δm for sensor m
Experimental results(1/3) • Frequency offs= 8 kHz and a cross-array with M = 5 microphones • 16 sets of 6 speech signals (3 male, 3 female, different for each of the 16 sets) • SNR was between 20 and 30 dB • Typical values are: fl = 250Hz,t2 = 20 dB, t3 = 0.2
Conclusion • Presented the NOSET algorithm to estimatethe number of sources in blind source separation. • It relies on DOAestimation at selected one-source TF points and works in both overdetermined andunderdetermined situations.