Adaptive 3D MV beamforming in medical ultrasound imaging

Abstract

In the recent years, volumetric imaging or 3D ultrasound imaging has been very attractive in comparison with the other diagnostic tools in medical imaging systems. In this paper, we used 2D planar array probe to obtain the RF data in order to construct the 3D image of a volume. The conventional method of beamforming in both 2D and 3D medical ultrasound imaging systems is delay and sum (DAS). We used a 2D Hanning window to simultaneously apply the Hanning window in x and y directions. DAS is a non-adaptive beamformer, then 3D adaptive minimum variance (MV) beamformer was proposed to apply to the simulated data derived from Field-II. In contrast to 3D conventional beamformer (3D DAS), adaptive 3D MV beamformer can significantly improve the imaging resolution of the simulated point targets. We implemented these beamformers in the MATLAB and survey the authenticity of them. The results illustrated that for a 32×32 elements probe with the central frequency of 3.5 MHz and sampling frequency of 20 MHz, FWHM of 3D MV beamformer for a point target is 9.36 times less than of 3D DAS beamformer and also side lobe level of 3D MV for a single point target, is 13.6dB lower than 3D DAS beamformer. It means that the resolution of 3D MV beamformer is better than 3D DAS. In order to compare the contrast of two types of beamformer, a cyst phantom was simulated in a speckle pattern. The results show that the proposed adaptive 3D MV beamformer significantly outperforms the 3D DAS.