肺水肿三维电阻抗断层成像方法

3D electrical impedance tomography method for pulmonary edema

为了对肺水肿过程进行连续可视化监测,实时监测肺水肿动态特征,提出了一种基于Split-Bregman算法的3D-TV正则化三维EIT成像方法。将3D-TV正则化方法与传统的共轭梯度(conjugate gradient,CG)算法和总变差正则化(total variation,TV)算法重建图像的相对误差(relative error,RE)、相关系数(correlation coeffi-cient,CC)和结构相似度(structural similarity index measurement,SSIM)进行对比与评价;并将3D-TV正则化方法应用于大鼠肺水肿实验,监测大鼠肺水肿的整个过程并进行分析。结果表明:与传统TV算法相比,3D-TV正则化方法使三维EIT成像的RE降低了9.1%,CC提高了9.1%,SSIM提高了7.7%,说明3D-TV正则化方法更适合三维EIT图像重建,并且可以对大鼠肺水肿的整个过程进行监测和成像。

In order to realize continuous visual monitoring of the pulmonary edema process,monitor the characteristics of dynamic changes of pulmonary edema in real time,a 3D-TV regularized 3D-EIT imaging method based on Split-Bregman algorithm is proposed.The 3D-TV regularization method is compared with the traditional conjugate gradient(CG)and total variation(TV)algorithms to evaluate the relative error(RE),correlation coefficient(CC)and structural similarity index measurement(SSIM)of the reconstructed images.Finally,the 3D-TV regularization method was applied to the rat pulmonary edema experiment,and the whole process of rat pulmonary edema was monitored and analyzed.The results show that compared with the traditional TV algorithm,the 3D-TV regularization method reduces the RE of 3D EIT imaging by 9.1%,improves CC by 9.1%and improves SSIM by 7.7%,indicating that the 3D-TV regularization method is more suitable for 3D-EIT image reconstruction.And the whole process of pulmonary edema in rats can be monitored and imaged.