simulated_head_sinograms.mat

Simulated head sinograms<br>Mats Persson, Jonas Adler<br>Latest version 2017-09-08<br><br>These are simulated basis materials sinograms used for the proceedings paper <br>M. Persson* and J. Adler*, “Spectral CT reconstruction with anti-correlated noise model and joint prior”, in Proceedings of The 14th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine, 2017, pp. 580–585.<br><br>This is meant as an example dataset for the reconstruction code belonging to this paper, which is available online at https://github.com/adler-j/spectral_ct_examples<br><br>Contents of the .mat file:<br>multiPhantom (600x600x2)<br>This contains the two basis images for the FORBILD head phantom: multiPhantom(:,:,1) is the soft tissue image and multiPhantom(:,:,2) is the bone image.<br><br>decomposedbasisProjectionsmm (853x360x2)<br>This matrix contains the basis sinograms (soft tissue and bone in that order) estimated from measured data using basis material decomposition. The values are given as projected path lengths in mm. The first index is the detector element number, the second index is view angle and the third image is the basis material.<br><br>CRLB (853x360x2x2)<br>This is the Cramer-Rao lower bound on basis image variance estimated from the noisy projection data for each projection line. The first two indices are detector element number and view angle. CRLB(i,j,:,:) is the lower bound (and in practice also an accurate estimate) of the covariance matrix of the estimated projected thicknesses of soft tissue and bone in ray (i,j). So CRLB(i,j,1,1) is the soft tissue image variance and CRLB(i,j,2,2) is the bone variance, CRLB(i,j,1,2)=CRLB(i,j,2,1) is the covariance. <br>The unit is mm^2.