Image-domain correction for gray level variation in circular cone-beam CT

It is well-known that the so-called gray level variation (mainly direct current (DC) shifting) across slices along z-direction, especially within the outer CT slices can be observed in circular cone-beam CT (CBCT). The underlying cause of the effects may partially attribute to the approximation characteristic of analytic reconstruction algorithm, either filtered backprojection (FBP) or backprojection-filtration (BPF) type, for circular CBCT imaging. To eliminate the DC shifting artifacts, helical scanning trajectory may be used for CBCT imaging, hence complete projection data being generated for an exact reconstruction according to data sufficiency condition. However, circular CBCT is still the most desirable for its inherent superiority in the convenience and stability of operation. So, an image-domain correction for the DC shifting is of its practical merit. In this work, using Forbild head mathematical phantom, we conduct a study to investigate the performance of the correction for the DC shifting by histogram-based clustering and spline-based interpolating. It is believed that the preliminary results presented in this paper can provide an effective path in the reduction of DC shifting, which has been annoying CT research community for a long period of time.