Heart Image Based On Displacement Field Snake Design And Implementation Of Segmentation System

According to related reports,heart disease has become the disease with the highest mortality rate,and the incidence rate is also continuously increasing.Heart disease is seriously threatening human health,so the prevention and diagnosis of heart disease is very urgent.With the development of Magnetic Resonance Imaging(MRI)and the non-invasive nature of MRI and the adaptation to soft tissue,MRI has been widely used in medicine,using NMR.Imaging technology accurately tracks the movement of the heart.Analysis of labeled cardiac MRI images helps doctors diagnose heart disease,find heart abnormalities more quickly,and locate abnormalities.Due to the complexity of the heart structure,and the effects of magnetic marker lines,papillary muscles,weak borders,and artifacts on heart edge information.The traditional segmentation model based on image grayscale and image texture features is not ideal for automatic heart segmentation of labeled cardiac magnetic resonance imaging(TMRI).It is difficult to achieve accurate segmentation.Therefore,the automatic segmentation of the heart has been a challenging research topic in the study of cardiac medical images.In order to solve the above problems affecting heart segmentation,this paper proposes a new method of whole heart segmentation.By observing the continuous sequence of labeled nuclear magnetic resonance images,we clearly found the frequency of motion of the heart region in cardiac MRI images.It is higher than other areas,so the heart’s motion information can be used to segment the heart.This paper is based on the motion characteristics of the heart region in the labeled cardiac magnetic resonance image.Based on the(Gradient Vecto Flow Flow Snake GVF Snake)model,a new method is proposed.The Snake model based on Displacement Gradient Vector Flow(DGVF Snake)is used to achieve automatic segmentation of the whole heart of labeled cardiac nuclear magnetic resonance images.In the first step,the displacement field of the adjacent two-frame labeled cardiac magnetic resonance image is obtained by the Modeling of Local Sinusoidal Waves(SinMod)method.The construction steps of the displacement field are as follows:(1)Designed by The bandpass filter filters the two cardiac nuclear magnetic resonance images,and the filtered image contains all the details of the heart motion.(2)The power image is calculated by the filtered image.(3)Calculate the local frequency and local phase shift by power image calculation to calculate the two-dimensional motion displacement field of the labeled two-frame nuclear magnetic resonance image;the second step,the displacement field and the original GVF field based on the SinMod method The weight field summation with the displacement field forms an external force field for constructing the DGVF,thereby realizing automatic segmentation of the heart by the DGVF Snake model.The system is based on OpenCV on the Qt Creator platform and implements a displacement-based heart Snake segmentation desktop application system to help doctors diagnose heart disease.In this paper,through the experiment and test of sequential tagged cardiac magnetic resonance images,it is proved that the proposed Snake mode segmentation method based on displacement field can segment the heart from tagged cardiac magnetic resonance images comparatively,and the segmentation results are more accurate.