Self-Diffusion Maps from Wavelet De-Noised NMR Images
A wavelet transform was used to improve the signal-to-noise in NMR microimages of Phaseolus vulgaris cotyledonary tissues. Two cases were tested: a relatively clean image and, at longer TE, a relatively noisy image. Significant improvement in the image quality of both test cases was achieved, with the most dramatic improvement noted in the noisy image. The conditions were extended to include diffusion weighting with the aim of computing diffusion maps for the tissues. Again, the wavelet procedure was effective in improving the appearance of the (diffusion-weighted) images. Finally, diffusion maps from the raw and de-noised data were computed and compared. The de-noised images produced maps with better detail and more consistent features. It is concluded that wavelet de-noising is a very useful approach for enhancing the image quality of hard-to-image tissues.