Automatic real-time MCE sequence image myocardial tissue region-of-interest tracking method

Abstract

The invention discloses an automatic real-time MCE sequence image myocardial tissue region-of-interest (ROI) tracking method and belongs to the medical ultra-sound sequence image analysis processing technology field. The method is characterized in that, a medical image elastic registration method based on hierarchical strategies is in combination with the Karman filtering technology to realize automatic real-time MCE sequence image myocardial tissue ROI tracking. The method comprises steps, the myocardial ROI is selected; initiation is carried out; the medical image elastic registration method based on the hierarchical strategies is employed for registration of sequence images; correction on the registration result is carried out by utilizing the Karman filtering technology; and the myocardial ROI is marked on the sequence image. Through the method, automatic real-time MCE sequence image myocardial tissue ROI tracking can be reliably realized, the tracking method has relatively high precision and accuracy, and inter-frame accumulation of a tracking error can be effectively avoided. The method is further applicable to automatic tracking on other medical ultra-sound sequence image ROI.

Description

Technical field [0001] The invention relates to an automatic tracking method of a region of interest (ROI) of medical ultrasound serial images, in particular to a method for automatically tracking a region of interest (ROI) of myocardial tissue in real-time myocardial contrast echocardiography (real-time MCE) serial images, belonging to medical ultrasound Sequence image analysis and processing technology field. Background technique [0002] Real-time myocardial contrast echocardiography (real-time MCE), as a new technology that can non-invasively and quantitatively assess myocardial microcirculation, is moving from basic research to clinical practice. However, in the clinical practice of imaging medicine, factors such as subtle changes in the position or orientation of the ultrasound probe, heartbeat, respiration, patient movement, organ peristalsis and its elastic deformation, etc. cannot be completely avoided, resulting in real-time MCE sequence images. The misalignment betwee...

AI Technical Summary

Problems solved by technology

Dr. Lope of the University of Campinas in Brazil proposed a semi-automatic MCE sequence image registration algorithm based on segmentation and pixel attribute-based image registration technology (MLLopes, et al, Improving myocardial contrast cardiocardiography images registration by increasing images similarity. WCU, 2003), the registration accuracy and effective The accuracy depends on the template manually selected in the segmentation stage, and the registration process cannot be automated

Sun Fengrong, Wang Xiaojing, Jia Xiaobo and others and their research teams have carried out related work in the past (Wang Xiaojing, Application Research of Speckle Tracking Technology in Echocardiography Image Analysis and Processing, Master Thesis of Shandong University, 2007; Jia Xiaobo, Echocardiography Applied Research on Analyzing and Processing Several Issues, Master Thesis of Shandong University, 2009; Sun Fengrong et al., A Method for Quantitative Analysis of Myocardial Acoustic Contrast Imaging, ZL200910020745.3, 2011; Li Xincai et al., Real-time Acoustic Myocardial Contrast Imaging Based on Speckle Tracking Technology Quantitative Analysis, Aerospace Medicine and Medical Engineering, 2010), during which the speckle tracking method based on block matching and optical flow field was used to solve the problem of automatic tracking of myocardial ROI in real-time MCE quantitative analysis, but real-time MCE sequence images are not very good The two basic assumptions of the block matching method (pixels in the same block have the same motion vector, and the pixels in the block are only translational motion) and the three basic assumptions of the optical flow field method (brightness invariance, gradient invariance and smoothness ), the tracking accuracy, accuracy and robustness of the method are not ideal, the inter-frame accumulation of tracking error is large, and the computational complexity of the speckle tracking method based on the optical flow field is relatively high

Afterwards, Sun Fengrong, Wang Li, Gao Xinjian and others respectively made medical image elastic registration technology (Wang Li, Application Research of Medical Image Registration Technology, Master Thesis of Shandong University, 2012) and Kalman filter technology (Gao Xinjian, real-time myocardial Acoustic Contrast-enhanced Quantitative Analysis of Several Image Processing Problems, Master Thesis of Shandong University, 2013) The application feasibility of the above-mentioned myocardial ROI automatic tracking problem was initially explored, but the respective application performances of the two technologies (especially the two However, the accuracy, robustness and practicality of each technology cannot meet the needs of clinical applications.

Images