Fig. 1.

Singular value decomposition (SVD). (a) The 3D matrix of complex-valued holograms $H(x,y,t)$ is reshaped into a 2D space-time matrix, and decomposed in a product of 3 matrices following Eq. (1). $U$ and $V$ are the spatial and temporal eigenvectors, and $\mathrm{\Delta }$ is the diagonal matrix of singular values ${\lambda }_i$. (b) Ordered singular values in dB. (c) Fourier transform magnitude of the temporal eigenvectors weighted by singular values, the arrow indicates high frequency clutter. (d) Individual or averaged spatial eigenvectors: the first vectors show clutter whereas vectors associated to singular values of lower energy reveal blood flow.