A Generally Regularized Inversion for NMR Applications and Beyond

Abstract

NMR (Nuclear Magnetic Resonance) technique presents a powerful analytical tool in the fields of chemistry, biology, and material science. The performance of the NMR technique highly relies on post-data processing methods, which are diverse according to different applications. In this paper, we propose a general processing method that can cover a wide range of applications from molecular dynamics study to biomedical engineering. We analyze the computation performance of the method by comparing it to other well-known algorithms and demonstrate its application in various NMR data processing problems, i.e., diffusion-ordered spectroscopy processing, under-sampled MRI (Magnetic Resonance Imaging) reconstruction, Laplace NMR inversion, and spectrum denoising. The results corroborate that the method performs effectively and efficiently compared to state-of-the-art methods. Moreover, we construct the method as a light toolbox, with which the users can perform various kinds of data processing tasks conveniently via graphic interfaces.