Low Complex Beam-Space Beamformer for Medical Ultrasound Imaging Based on Delay Multiply and Sum Method

Abstract

The beam-space minimum variance (BSMV) beamforming is an outstanding form of beam-space beamforming. However, its computational complexity is still high for real-time ultrasound imaging. To solve this problem, a beamforming named beam-space delay multiply and sum (BS-DMAS) beamforming was proposed in this paper. In the proposed method, similar decimation discrete cosine transform (DCT) matrix in BSMV was used to reduce the dimension of signal firstly. Then, the improved delay multiply and sum (DMAS) beamforming in parallel form was used to get the final output. Different from BSMV, instead of calculating the inverse of covariance matrix to get the adaptive weight vector, the proposed method just need to perform square root, sign and addition operation. By doing this, the computational complexity was reduced from PM + P3 + 3P2 + O(P) to PM + P2 + O(P) compared with BSMV. Finally, simulation and experiment was performed to evaluate the proposed method. Compared with DAS, the results indicate that BS-DMAS generates a 66% lower advantage at full width at half-maximum (FWHM), an 114% improvement at contrast noise ratio (CR) but slightly worse at contrast ratio (CNR). This demonstrates that the proposed method can improve the performance of imaging and reduce the computational complexity.