Modeling and validation for compressed sensing and MRI

摘要

A method judiciously applies or manages randomness, incoherence, nonlinearity and structures for improving signal encoding or decoding. The method in a compressed sensing-based imaging example comprises acquiring a set of base data samples, obtaining a base result, perturbing the base set, obtaining perturbed result(s), and deriving an outcome facilitating assessment and improvement of image quality. The method in a magnetic resonance imaging example comprises acquiring data samples in parallel and in accordance with a k-space sampling pattern, identifying a signal structure in an assembly of the acquired data samples, and finding a result consistent with both the acquired data samples and the identified signal structure.