These are two publications that measure the perceptual consequences of changes in reconstruction kernel. Early in development of CT, I had an interest in two reconstruction procedures. (Shepp-Logan and Ram-Lak). These were slightly different back-projection convolution procedures. Given my interest in perception, the difference in visibility of features in CT images was an obvious question. Setting up this experiments described in these papers, we realized that choice of reconstruction kernel was linked with display of the reconstructed image. Narrow windows benefitted the display of images using smooth kernel. However, narrow windows created thresholding issues. The 1985 paper was published in British Journal of Radiology. The journal choice was prompted by a paper by Warren that described a window effect in British Journal of Radiology. |
Home page > Accuracy of CT Numbers >