基于结构张量全变差的CT图像去噪算法

CT Image Denoising Algorithm Based on Total Variation of Structure Tensor

在计算机断层成像过程中,为了降低辐射剂量,往往采用稀疏角度投影数据,滤波反投影重建算法(Filtered BackProjection,FBP)具有重建速度快的优点,但在稀疏角度情况下重建得到的图像中存在较严重的条状伪影。针对此问题,该文研究采用基于结构张量全变差(The Structure tensor Total Variation)CT图像后处理去噪算法,该文中简称该算法为FBP-STV算法。为了验证该算法的可靠性,该文用2类模型进行数据仿真实验,并与FBP算法、截断全变分2类算法进行比较,实验结果表明,该文研究的FBP-STV算法,去噪效果相对较好,图像质量相对较高。

In the process of computed tomography,sparse angle projection data is often used to reduce radiation dose.Filtered back projection(FBP)reconstruction algorithm has the advantage of fast reconstruction speed,but there are serious strip artifacts in the reconstructed images in the case of sparse angles.In order to solve this problem,this paper studies the denoising algorithm of CT image post-processing based on structure tensor total variation.In this paper,the algorithm is called FBP-STV algorithm for short.In order to verify the reliability of the algorithm,data simulation experiments are carried out with two kinds of models and compared with FBP algorithm and truncated total variational algorithm.The experimental results show that the FBP-STV algorithm studied in this paper has a relatively good denoising effect and relatively high quality of images.