Integration of compressed sensing and frequency hopping techniques for ultrasound tomography

Abstract

The FH-DBIM (frequency hopping distorted Born iterative method) is a promising approach for imaging tumors at the level of biological tissues. This uses multiple frequencies for producing images, leading to large computational complexity and storage space (i.e. for each frequency, NtNrN2 calculations are required). It is a main barrier that the FH-DBIM approach is not fully exploited in commercialized devices. Therefore, in this paper, we propose an alternative algorithm for significant reducing the computational complexity of the FH-DBIM by integrating compressed sensing and frequency hopping techniques for the DBIM. That is, the needed number of measurements for imaging is tremendously reduced in comparison with the conventional one. The proposed algorithm - CS-FH-DBIM-offers a new prospect for being able to image high-resolution images at the level of biological tissues in practice.