Sinogram Correction Methods Using Sinogram Decomposition

摘要

We suggest a novel approach, called the sinogram decomposition, that can be used for sinogram correction. One application of this approach is the outer problem, when the outer parts of the scanned object extend outside the measured field of view, but still need to be reconstructed. Another application is the internal problem, when the data in central part is corrupted by high noise and/or artifacts. In both cases, sinogram curves have parts that are valid, that is, measured and not corrupted. We use these valid parts for estimation of the missing, or corrupted, parts of the sinogram; each estimated sinogram pixel is obtained by integrating over the family of the sinogram curves through this pixel.