Practical hybrid convolution algorithm for helical CT reconstruction

Great strides have been taken in the last few years in the development of both approximate and exact reconstruction algorithms for helical cone-beam computed tomography (CT). However, it is hard to achieve a good balance between reconstruction speed, flexibility, and image quality. We propose a new algorithm that combines the advantages of many previously published algorithms. It uses the so-called hybrid convolution, which is the sum of the ramp and Hilbert filters. In this work, we evaluate the new algorithm and compare it to other candidates in terms of spatial resolution, noise, and image artifacts. Our evaluation demonstrated that the proposed algorithm outperforms the helical Feldkamp algorithm in terms of image noise uniformity and the cone beam artifact. We also propose a simplified version for the over-scan reconstruction.