Dependent Gaussian mixture models for source separation

EURASIP Journal on Advances in Signal Processing

Table 2 Parameter estimated mean for WMAP data

a = 1	${\hat{p}}_{1} = 0.12$	${\hat{μ}}_{1} = (\begin{matrix} - 0.012 \\ 0.051 \\ - 0.001 \\ 0.017 \end{matrix})$	${\hat{Q}}_{1} = (\begin{matrix} 5.29 0 0 0 \\ 0 1.47 - 0.04 - 1.60 \\ 0 - 0.04 0.41 0.32 \\ 0 - 1.60 0.32 3.79 \end{matrix}) \times 1 0^{5}$
a = 2	${\hat{p}}_{2} = 0.88$	${\hat{μ}}_{2} = (\begin{matrix} - 0.012 \\ 0.029 \\ 0.003 \\ 0.017 \end{matrix})$	${\hat{Q}}_{2} = (\begin{matrix} 1.45 0 0 0 \\ 0 0.54 0.08 - 0.20 \\ 0 0.08 0.28 0.26 \\ 0 - 0.20 0.26 1.09 \end{matrix}) \times 1 0^{4}$