1) To appreciate that substantial preprocessing of the measured data takes place in CT prior to image reconstruction. 2) To understand that this preprocessing serves to correct for degradations in the data that would lead to artifacts in reconstructed images if not accounted for. 3) To give a sense of the current approaches used for this preprocessing as well as some emerging new approaches.
While most medical image processing is performed on the familiar anatomical and physiological images presented to radiologists, tomographic imaging modalities also involve a large amount of image preprocessing that is performed on the raw data acquired by the scanners. This preprocessing takes place prior to the image reconstruction step necessary to convert the raw data into clinically useful images and serves to correct the data for physical degradations. These corrections bring the data into accordance with the somewhat idealized imaging models that underly most image reconstruction algorithms; failure to perform these corrections would
generally lead to unacceptable artifacts in reconstructed images. In addition to mitigating artifacts, these preprocessing steps can also be used to tune the resolution and noise properties of the final reconstructed images.
Data preprocessing prior to reconstruction is particularly important in computed tomography (CT) because the very high contrast resolution of the reconstructed images and the narrow display grayscales employed means that even small inconsistencies in the raw data can lead to distracting artifacts in reconstructed images. In this chapter, we describe a number of the most common CT data degradations and the data preprocessing strategies that are used to correct or mitigate them.
La Riviere, P,
Sinogram Preprocessing for Artifact Reduction in CT. Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL.
http://archive.rsna.org/2007/4405281.html
