乳腺肿瘤同伦LM算法EIT图像重建  

Image Reconstruction of Breast Tumors Using Homotopy LM Algorithm in Electrical Impedance Tomography

作　　者：秦伟刚[1,2] 王超[1] 

机构地区：[1]天津大学电气与自动化工程学院,天津300072 [2]天津工业大学电气工程与自动化学院,天津300387

出　　处：《天津大学学报（自然科学与工程技术版）》2016年第5期506-512,共7页Journal of Tianjin University：Science and Technology

基　　金：国家自然科学基金重点资助项目(50937005)

摘　　要：将同伦延拓方法与Levenberg-Marquardt(LM)迭代算法相结合,对不同的乳腺肿瘤模型进行图像重建,并与LM算法进行比较.在不同迭代初始值下,同伦LM算法与LM算法的平均电导率相对误差相比,最优值能降低接近7%,,运行时间最大能减少46%,的计算开销,电压相对误差要优于LM算法7%,以上.实验结果表明,同伦LM算法比LM算法能重建更好的图像,提高了重建图像的精度和收敛速度.A homotopy Levenberg-Marquardt（HLM）algorithm,which is the combination of homotopy continuation method and Levenberg-Marquardt（LM）algorithm,was presented.The reconstructed images of different breast tumor models were acquired by HLM algorithm and compared with those by LM algorithm.The relative errors of the average conductivity were obtained by means of HLM algorithm and LM algorithm.The optimal value of the relative errors of the average conductivity by HLM algorithm was reduced by nearly 7 percent for different initial values.The best running time and the relative errors of the voltage are decreased by 46 percent and 7 percent respectively.The results show that the better images are reconstructed by HLM algorithm,and the accuracy and convergence rate are improved.

关 键 词：电阻抗层析成像 同伦延拓法 LEVENBERG-MARQUARDT算法 

分 类 号：R318[医药卫生—生物医学工程]