Article ID Journal Published Year Pages File Type
411466 Neurocomputing 2016 6 Pages PDF
Abstract

The first generation of omni-directional M-mode echocardiography system is able to extract implied motion information from sequential echocardiography images. However, in recent years, there has been an increasing demand in clinical practice for more precise dynamic motion information and research into the second-generation of omni-directional M-mode echocardiography systems has focused on how to achieve this greater precision. Two possible approaches are to improve the acquisition precision of echocardiography and to improve the extraction accuracy of motion curves. Firstly, we describe the use of a model for separation of cardiac ‘non-functional’ movement, in the design of a particle swarm optimization (PSO) algorithm to track and analyze the heart feature points, in order to achieve improved omni-directional M-mode echocardiography using tracking sampling lines. We then present the design of a new method for heart motion curve extraction, based on the idea of multi-scale analysis and wavelet transformation, which can suppress image noise effectively and thereby reduce the need for manual intervention. Experimental evaluation indicated that these techniques produce more effective results than the first generation omni-directional M-mode echocardiography system.

Related Topics
Physical Sciences and Engineering Computer Science Artificial Intelligence
Authors
, , , ,