Resolution improvement in phantom and in vivo through l1 regularized coherent compounding

Compressive Beamforming uses a beamforming matrix in conjunction with a sparsity prior to form high quality ultrasound images from a single or few insonifications. In some cases, due to poor SNR or model mismatch, it is desirable to use a larger number of transmit beams. We extend our framework to introduce coherent compounding as a way to process a large number of transmits without increasing memory requirements. It leads to reduced artifacts and improved resolution and contrast in phantoms and in vivo. We also apply this formalism to focused beams.