Modeling element directivity in minimum variance beamforming for medical ultrasound

Minimum variance (MV) adaptive beamforming, which adaptively estimates and suppresses interfering signals, has attracted increased attention for ultrasound imaging in recent years. While many MV algorithms have been applied to ultrasound, these algorithms generally rely on signal models that neglect amplitude differences across the array due to element directivity or other factors, introducing mismatch for near-array sources. This paper proposes a beamspace method using randomly sampled subarrays which allows the use of higher-fidelity signal models. Using experimental data, we demonstrate that this algorithm yields noticeable gains in signal contrast by accounting for element directivity.