Spatial and Temporal Adaptive FIR Clutter Filtering

In this study, a new spatial and temporal FIR clutter filter has been developed using the tissue velocity. This filter uses the PSF attenuation and the tissue velocity estimates to design a filter which is time-dependent but also a function of the position relative to the tissue. The spatial and temporal adaptive FIR filter is compared with three standard FIR clutter filters: (i) a low cutoff FIR filter, (ii) a high cutoff FIR filter, and (iii) a temporal adaptive FIR filter. In the following, this comparison was made with an in vivo carotid data set from a high frame rate acquisition using unsteered plane waves. The axial component of tissue velocities was estimated using a 1D phase-based estimator. For blood flow, 2D vector velocities were estimated using a 2D phase-based estimator after the introduction of transverse oscillation in post-acquisition. Overall, both adaptive filters outperform the non-adaptive filters. Indeed, standard filtering methods yield strong outliers in the presence of low flow velocity for the maximum filter and high flow velocity for the minimum filter. Because the spatial and temporal adaptive filter takes into account the local tissue velocity, it facilitates flow velocity estimation closer to the wall, particularly for low flow. Indeed, only the relevant (based on the PSF) tissue velocity is removed from a specific location in the lumen. Consequently, a local high tissue velocity does not influence the cutoff velocities in the entire vessel. These new adaptive filtering methods could lead to better separation between blood and tissue in recordings from pathological carotids with spatially varying tissue velocity.