Constraint Matrix Projection Minimum Variance Beamformer Based on Adaptive Decomposition Threshold for Ultrasound Imaging

The eigenspace-based minimum variance (ESBMV) beamformer can enhance the contrast of the minimum variance (MV), but the traditional orthogonal projection and fixed eigen-decomposition threshold make it difficult to obtain further improvement in resolution and speckle uniformity. In this paper, a constraint matrix projection minimum variance based on adaptive decomposition threshold (ADCPMV) is proposed for ultrasound imaging, which aims to improve the insufficiency of ESBMV. Firstly, a constraint matrix with direction information is put forward to better remove the interference and noise of MV weight. Then, a threshold decomposition function based on truncated smoothed coherence factor (TSF) is proposed to eliminate dark-area artifacts. Finally, the TSF is used as the weighting to improve the ADCPMV contrast. Simulations and the experimental datasets of ats_wires, geabr_0 and rat_tumor are introduced to inspect the performance of the proposed algorithms. The result of point target simulation shows that the full-width at half-maximum (FWHM) of ADCPMV-TSF is reduced by 93.98% and 80.19% than that of traditional delay-and-sum (DAS) and ESBMV respectively. Furthermore, the geabr_0 experimental result shows that compared with traditional DAS and ESBMV, the FWHM of ADCPMV-TSF is improved by 87.74% and 61.19% respectively. Meanwhile, the proposed ADCPMV-TSF can well avoid the generation of dark-area artifacts and get the maximal improvement of speckle signal-to-noise ratio is 103.57%.