Adaptive Multiple Snapshots with Matrix Pencil Method for Color Flow Imaging

Matrix Pencil (MP) estimation framework, which uses a generalized eigenvalue problem formulation and matrix rank reduction principle, has been recently introduced for color flow imaging. Based on an adaptive rank selection method, MP can extract flow velocity directly from raw slow-time ensembles whose clutter has not been suppressed. Therefore, biases due to clutter filter can be avoided, and MP is more accurate than conventional lag-one autocorrelation method. This paper attempts to make use of multiple snapshots for fast-time averaging, which removes aliasing caused by noise and provides precise estimation in low signal-to-noise ratio (SNR) blood flow environment. Further, an adaptive method is proposed for choosing averaging window size adaptively based on different SNR and tissue stationarity. Analysis based on synthesized flow data shows that noise can be effectively suppressed, and velocity resolution is preserved simultaneously.