A globally constant sound speed of 1540 m/s is often assumed when beamforming ultrasound images, though variations in tissue properties cause local sound speed differences, degrading image quality with inaccurate target depths and shapes, thus reducing the overall effectiveness of diagnostic and interventional ultrasound imaging. In this paper, we developed a more general coherence-based optimization method to estimate sound speed, applicable to both coherent and incoherent targets in plane wave ultrasound imaging. Sound speeds that produced minimum values of the lateral full-width at half maximum (FWHM) of point targets and the contrast of anechoic targets were considered optimal. Our method estimated sound speeds that deviated 5-10 m/s from optimal values. Our approach was benchmarked against a speckle brightness maximization approach, which deviated 57-160 m/s from optimal values. The lower deviations with our approach resulted in a 46%-63% FWHM improvement and a 1.09 dB contrast enhancement relative to the corresponding results of speckle brightness maximization. The proposed approach shows promise for improving accuracy in diagnostic and interventional ultrasound imaging.
