Compressed sensing for beamformed ultrasound computed tomography

Ultrasound computed tomography (UCT) allows reconstruction of quantitative tissue characteristics. Lowering the acquisition time would be beneficial; however, this is limited by the time of flight and the number of transmission events. Moreover, corruption of the measurements by noise may cause inverse scattering reconstruction methods such as the Born Iterative Method (BIM) to converge to a wrong solution. Beamforming using multiple elements to obtain a narrow beam has the potential to mitigate the effects of noise; however, spatial coverage per transmission event reduces in this case. To excite the full domain, more transmissions are required and the acquisition time increases even further. We therefore consider compressive acquisitions based on parallel randomized transmissions from a circular array. Relying on the assumption that the object is compressible, we combine the BIM with sparse reconstruction to obtain the estimated image.