부산대학교 도서관

The emergence of ultrasound plane wave technology has driven the development of ultrasonic ultrafast frame rate imaging technology. However, simultaneously, it has placed an increased burden on hardware sampling, transmission, and data storage. Compressed sensing technology is widely employed in ultrasound sparse sampling and imaging. However, it faces challenges in effectively recovering useful signals when the sampling rate is low or the echo noise is excessive. Based on the characteristics of ultrasound echo signals, this paper employs a multi-hypothesis prediction method to reconstruct correlated echo signals. This approach can efficiently reconstruct ultrasound echo signals while reducing the reliance on the sparsity of the original echo signals. On PICMUS experimental data, the image similarity surpasses that of methods relying solely on sparsity or both sparsity and correlation. Without compromising image quality, this paper's method accomplishes signal compression and reconstruction from the perspective of signal correlation.