3D spatial resolution characterization for volumetric computed tomography

Abstract

Spatial resolution is a key parameter that determines the smallest resolved scale in volumetric computed tomography. However, the evaluation process is typically confined within a plane, and tomographic resolution is usually treated as a two dimensional concept. In this work, the full map of three-dimensional (3D) spatial resolution along radial directions was obtained through combined volumetric laser induced fluorescence measurement on a well-controlled test rig, tomographic reconstruction, and successive resolution analysis. The results show non-uniformly distributed spatial resolution and weak dependence on view angles. Moreover, both experimental and algorithmic factors on spatial resolution were quantitatively quantified. The convergence of the spatial resolution was then validated through a series of computation with increasing voxel numbers. It is found that the spatial resolution improves until the number of voxels reaches ∼19.6 times as much the value set by the linear algebraic limit. The ultimate spatial resolution that can be obtained with our setup is about 1.35 lp/mm, which is twice the value of a well-posed system. The results show that the method of this work is more appropriate to quantify spatial resolution in 3D space. The conclusions made in this work can be generalized for applications with other tomography-based measurements.