去噪自编码器在电阻抗深度成像中的应用  

Application of denoising autoencoder in electrical impedance depth imaging

作　　者：韩仲鑫 刘瑞兰[1] 左瑞雪 钱慧 HAN Zhongxin;LIU Ruilan;ZUO Ruixue;QIAN Hui(College of Automation&College of Artificial Intelligence,Nanjing University of Posts and Telecommunications,Nanjing 210046,China)

机构地区：[1]南京邮电大学自动化学院人工智能学院,江苏南京210046

出　　处：《现代电子技术》2023年第20期55-60,共6页Modern Electronics Technique

摘　　要：电阻抗成像技术是一种安全无创的医学成像技术,且没有辐射带来的危害,但其逆问题是一个高度不适定的病态非线性问题,获得的重构图像伪影较大。为了得到较为清楚的重建图像,提出一种将传统方法与去噪自编码器相结合的深度成像方法。该方法首先使用分裂Bregman方法粗略重建图像,之后采用去噪自编码器对粗略重建图像进行去噪处理。仿真和实测结果表明,针对圆形目标,所提方法能够实现较高精度地去除伪影和精确的形状重建,运算速度快,效果较好。Electrical impedance technology(EIT)is a safe and non⁃invasive medical imaging technology and there is no hazard caused by radiation.But inverse problems are a highly unsaturated pathological nonlinear problem,and the reconstructed image artifacts are large.In order to get clearer reconstructed image,a depth imaging method combining traditional method and denoising autoencoder(DAE)is proposed.In this method,the split Bregman method(SBM)is used to roughly reconstruct the image.The denoiding process for the roughly reconstructed image is conducted by means of the DAE.The simulation and experimental results show that the proposed method can achieve high⁃accuracy removing artifacts and precise shape reconstruction for circular targets,with fast computational speed and good results.

关 键 词：电阻抗深度成像 去噪自编码器 EIT模型 图像重建 SBM算法 结构相似性 

分 类 号：TN911.73-34[电子电信—通信与信息系统] R318[电子电信—信息与通信工程]