CT图像重建中的一种新型滤波器

滤波反投影图像重建算法中滤波器设计是其中至关重要的一环。针对现有一次指数滤波器空域波形存在明显振荡、频域曲线截止频率附近没有明显压低的问题，引入高斯滤波器减小Gibbs效应，并设计一种用于CT重建领域的新型滤波器，即改进高斯滤波器。通过仿真实验对比说明滤波器中加权因子选择对图像质量的影响，另与RL滤波器、SL滤波器、一次指数滤波器、高斯滤波器实验对比验证了该滤波器拥有良好的抗噪性能，有效减小重建图像灰度波动以及明显提高重建图像质量。

A novel filter for CT image reconstruction

Abstract: Filter design is an important part in the filtered back projection image reconstruction algorithm. In view of the problem that the existing first-order exponential filter has obvious oscillations in the space domain waveform and no significant pressure drop near the cut-off frequency of the frequency-domain curve, a Gaussian filter is introduced to reduce Gibbs effect, and a novel filter (improved Gaussian filter) is designed for CT image reconstruction. The comparison of simulation experiments reveals the effect of the selection of weighting factors in the filter on the image quality. In addition, the comparison among RL filter, SL filter, first-order exponential filter and Gaussian filter verify that the proposed filter has a good anti-noise performance, and that it can effectively reduce the grayscale fluctuation of the reconstructed image and significantly improve the quality of the reconstructed image.