Fast scatter correction for cone-beam computed tomography using the statistical method

X-ray scatter signals acquired from a large flat panel detector in cone-beam computed tomography (CBCT) can degrade image quality of cross-section images. In this paper, we propose a new fast scatter correction method in CBCT using the statistical method in combination with the downsampling technique in projection images to reduce the computation time. The downsampled primary signal is estimated using the MLEM algorithm and then upsampled to obtain the original size. To reduce the upsampling effect, we deconvolve one more time to obtain the final estimated primary signal. The experimental results show that the proposed method with downsampling by two can reduce the computation time by a factor of more than four. In addition, the quality of reconstructed images is improved in terms of higher CT number accuracy and smaller cupping artifacts in comparison with the results obtained from fan-beam computed tomography.