A pulse-echo sound speed estimation approach with prior constraints for layered media

This paper presents a pulse-echo sound speed estimation method for layered media, utilizing prior-constrained coherent analysis. The proposed method addresses the instability in local sound speed estimation caused by phase ambiguities resulting from suboptimal probe configurations. By introducing biologically reasonable sound speed boundary constraints to compensate for errors in average sound speed (ASS) estimation, and integrating sparse interface regularization inversion models, this method suppresses noise amplification during inversion, thereby enhancing robustness. The experimental results demonstrate that using this method significantly improves performance in simulations and in vitro data, reducing the root-mean-square error (RMSE) by 68% compared to existing methods. In in vivo experiments, the average sound speed in the tested regions deviated less than 0.6% from the reference values, while maintaining high repeatability. Furthermore, ablation studies validate the synergistic effect of prior compensation and sparse regularization, confirming their effectiveness in reducing phase sensitivity and enhancing the resolution of stratified structures. This method provides a reliable quantitative sound speed assessment tool for clinical scenarios such as hepatic steatosis, simultaneously relaxing hardware requirements for ultrasound probe parameters.