Mathematical Biology Seminar
This talk will provide an overview of structured low-rank algorithms and model based deep learning methods, with applications to MR imaging. I will briefly introduce structured low-rank algorithms for super-resolution, blind channel equalization, and inverse problems. The talk will then focus on non-linear generalizations of structured low-rank methods, and establish the links to kernel regression and deep learning. I will then discuss the integration of deep learning networks with model-based image recovery schemes. Various deep network architectures inspired by structured low-rank algorithms, with applications to recovery from undersampled MRI data and multichannel image recovery will be demonstrated.