Sharpening of CT images by cubic interpolation using B-spline

Computed tomography (CT) is a method by which an original image is reconstructed based on the data of projected images collected from various directions, and is widely used in the fields of medicine and industry. However, the edge of the image reconstructed by CT is unsharp, therefore, microflaws are often overlooked. We demonstrated that one of the causes of the decreased sharpness of the reconstructed image is associated with the linear interpolation during the back-projection process, and in our method, the linear interpolation is replaced by cubic interpolation using the B-spline. In addition, by calculating the control points of B-spline by Fourier transform, the process required for the calculation of the control points appears to be eliminated. In the experiment, the reduction of unsharpness to 1/2 that of the conventional method and the reduction of processing time to a level equivalent to that of the conventional method have been achieved.