A novel method for resolution of an overlapping nuclear magnetic resonance
(NMR) spectrum using the wavelet transform (WT) is proposed, one to the loc
alization character in terms of both space and scale of the wavelet transfo
rm, an NMR spectrum can be decomposed into a series of localized contributi
ons (details and approximations) at different resolution levels, which repr
esent the spectral information of different resolutions. With amplification
of the contributions at line resolution level and then reconstruction (inv
erse transform), the resolution of the reconstructed NMR spectrum a ill inc
rease, Therefore, the resolved spectrum can he obtained from a low-resoluti
on spectrum or an overlapping spectrum, Simulated data sets and a spectrum
of a biological sample (gramicidin-S) were investigated by this method. It
was proven that resolution of NMR spectra can be greatly improved by this a
pproach.