Three-dimensional weighting reconstruction algorithm for cone-beam CT based on ellipsoidal bounding box

Improving the quality of cone-beam CT imaging under large cone angle scan has been an important research direction of CT imaging technology. This paper analyzes the characteristics of missing data along the axial direction in Radon space under circular scanning trajectory and the intensity drop rule of FDK algorithm. A three-dimensional (3D) weighting reconstruction algorithm is proposed for cone-beam CT based on ellipsoidal bounding box. The proposed method does not need any operation of rebinning and only uses projection data to gain the minimum bounding box of reconstructed objects. Three-dimensional weighting function can be automatically generated according to the size and spatial position of its inscribed ellipsoid, and the weighting function is added into the 3D back-projection process to achieve the algorithm. With little increase in computation, this algorithm does not introduce other artifacts and is equivalent to FDK algorithm in the mid-plane. Simulation results show that the proposed algorithm significantly improves the accuracy of FDK algorithm, and reduces the cone angle artifacts.