Blind separation of electron paramagnetic resonance signals using diversity minimization
This paper presents a method for the blind separation of Electron Paramagnetic Resonance (EPR) spectroscopy signals that can aid in the detection of free radicals in living organisms. Observed EPR signals are often mixtures of source signals that are approximately ''sparse'', with a small number of narrow segments of the signal much larger than the remaining parts. We develop a method to separate the sources through minimizing a p-norm-like diversity measure under some mild assumptions which are generally valid for EPR signals. Simulations demonstrate that the proposed method performs well on EPR signal separation, with better robustness to noise compared to other techniques.
Copyright © 2010 Elsevier Ltd. All rights reserved.
Version: za2963e q8zaa q8zbd q8zc9 q8zde q8ze3 q8zf5 q8zg2