The Relationship Between Grey Levels And Hounsfield Units In Cone Beam CT ( CBCT ) Scanners

Good Essays

Studies in the mentioned disciplines can be broadly divided into two levels namely physical and computational level. Due to the cost of physical development, the tendency for the computational aspects has been raised. Technical parameters of scanners include Cycle time, spatial resolution, low-contrast resolution, uniformity, linearity, slice thickness, computed tomography dose index (CTDI), and pitch \cite{duan2017computed}. Mah et al. \cite{mah2010deriving} investigate the relationship between grey levels and Hounsfield units (HU) in cone beam CT (CBCT) scanners. To do so, a phantom of 8 different materials was created and imaged with different CBCT scanners. The phantom was scanned under three conditions which were (1) a sole phantom; …show more content…

The projection sub-systems of CT scanners has experienced changes in 3 aspects including parallel, fan, and cone beam systems \cite{sidky2008image} - \cite{vo2014reliable}. Sidky and Pan \cite{sidky2008image} proposed a theoretical framework to show how accurate circular cone-beam CT image reconstruction can be done from reduced data sampling. In data acquisition phase, they assumed that the estimated projection data is within a specified tolerance of the available data and the values of voxels are non-negative. To observe these constraints, a projection onto convex sets was considered, and the total variation (TV) gets minimized by steepest descent with an adaptive step-size. They argued that the TV algorithm can resolve low-contrast structures in the presence of high-contrast objects even when the projection data sets are limited in angular range or view number, and when the low-contrast object is not at the mid-plane of the circular x-ray trajectory. Qi and Chen \cite{zhihua2008direct} considered image reconstruction in fan-beam based CT in which the derived formula applies to the real domain. First, inverse Fourier transform was applied to use the multiplication instead of convolution in Fourier space. Afterward, the range of calculation was extended to cover redundancy. This method is subject to observing full circle scan in data acquisition mode.

Get Access