Space-constrained optimized Tikhonov regularization method for 3D hemorrhage reconstruction by open magnetic induction tomography.

Autor:	Chen Y; Tianjin Key Laboratory of Process Measurement and Control, School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, People's Republic of China., Dong F; Tianjin Key Laboratory of Process Measurement and Control, School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, People's Republic of China., Tan C; Tianjin Key Laboratory of Process Measurement and Control, School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, People's Republic of China.
Jazyk:	angličtina
Zdroj:	Physics in medicine and biology [Phys Med Biol] 2022 Nov 16; Vol. 67 (22). Date of Electronic Publication: 2022 Nov 16.
DOI:	10.1088/1361-6560/ac9e42
Abstrakt:	Objective . Open magnetic induction tomography (MIT) is a promising technique for detecting the intracranial hemorrhage due to the non-radioactive, non-invasive and portable features. However, severe inhomogeneity of the sensitivity distribution under the open MIT sensor array and the ill-conditioned nature of MIT inverse problem limit the imaging quality in hemorrhage reconstruction. More accurate and robust imaging algorithms are urgently needed in clinical diagnosis. Approach. In this study, the space-constrained optimized Tikhonov regularization (SOTR) method is proposed for 3D hemorrhage reconstruction by open MIT. The sensitivity matrix is optimized according to the characteristics of sensitivity distribution under the open MIT sensor array. To test the performance of the SOTR method, 3D anatomical head models with hemorrhages in different volumes and locations were established. The images of the hemorrhages were reconstructed by the Tikhonov regularization (TR), total variation (TV) regularization, isotropic SOTR, and anisotropic SOTR method. Correlation coefficientCC,localization errorLE,and volume errorVEwere calculated to evaluate the hemorrhage imaging quality. Main results . Compared with the traditional sensitivity matrix, the optimized sensitivity matrix has smaller column number and better uniformity, which alleviates the under-determined and ill-conditioned problem of MIT. The imaging results indicate that both the isotropic and anisotropic SOTR methods can effectively improve the reconstruction accuracy for the location and volume of the hemorrhages. Moreover, compared with the TR and TV methods, the two SOTR methods are more robust against the measurement noise. Significance . The proposed method improves the imaging quality of the intracranial hemorrhage, which promotes the clinical applications of open MIT. (© 2022 Institute of Physics and Engineering in Medicine.)
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu