Deep neural networks for ultrasound beamforming

In the past, members of our group developed a model based beamforming method called aperture domain model image reconstruction (ADMIRE). They tuned ADMIRE to improve ultrasound image quality by suppressing sources of image degradation such as off-axis scattering, reverberation, and phase aberration. In addition, the development of ADMIRE demonstrated that beamforming could be posed as a regularized nonlinear regression problem, which suggests that a deep neural network (DNN) might be used to accomplish the same task. Compared to regularized regression methods, DNNs are fast, adaptive, and fault tolerant.