Literature DB >> 7874417

Reconstruction of images from radiofrequency electron paramagnetic resonance spectra.

C M Smith1, A D Stevens.   

Abstract

This paper discusses methods for obtaining image reconstructions from electron paramagnetic resonance (EPR) spectra which constitute object projections. An automatic baselining technique is described which treats each spectrum consistently; rotating the non-horizontal baselines which are caused by stray magnetic effects onto the horizontal axis. The convolved backprojection method is described for both two- and three-dimensional reconstruction and the effect of cut-off frequency on the reconstruction is illustrated. A slower, indirect, iterative method, which does a non-linear fit to the projection data, is shown to give a far smoother reconstructed image when the method of maximum entropy is used to determine the value of the final residual sum of squares. Although this requires more computing time than the convolved backprojection method, it is more flexible and overcomes the problem of numerical instability encountered in deconvolution. Images from phantom samples in vitro are discussed. The spectral data for these have been accumulated quickly and have a low signal-to-noise ratio. The results show that as few as 16 spectra can still be processed to give an image. Artifacts in the image due to a small number of projections using the convolved backprojection reconstruction method can be removed by applying a threshold, i.e. only plotting contours higher than a given value. These artifacts are not present in an image which has been reconstructed by the maximum entropy technique. At present these techniques are being applied directly to in vivo studies.

Mesh:

Year:  1994        PMID: 7874417     DOI: 10.1259/0007-1285-67-804-1186

Source DB:  PubMed          Journal:  Br J Radiol        ISSN: 0007-1285            Impact factor:   3.039


  1 in total

1.  Comparison of maximum entropy and filtered back-projection methods to reconstruct rapid-scan EPR images.

Authors:  Mark Tseitlin; Amarjot Dhami; Sandra S Eaton; Gareth R Eaton
Journal:  J Magn Reson       Date:  2006-10-27       Impact factor: 2.229

  1 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.