Deep learning-based ejection fraction calculation method and system

Abstract

The invention provides a deep learning-based ejection fraction calculation method and system, and relates to the technical field of deep learning. The method comprises the steps: carrying out heart left ventricle segmentation on heart body data through a pre-trained neural network segmentation model to obtain corresponding left ventricle segmentation marker data; performing binarization processingon the left ventricle segmentation marker data to obtain left ventricle binarization body data with first voxel values and second voxel values; counting the voxel quantity sum of the first voxel values in the left ventricle binarization body data, and storing the voxel quantity sum as the left ventricle volume corresponding to the heart body data into a pre-generated volume queue; and after all frames of heart body data of the continuous heart body data are processed, respectively extracting the maximum value and the minimum value of the left ventricular volume stored in the volume queue, andperforming calculation to obtain the ejection fraction of the heart part. The method has the beneficial effects of effectively improving the calculation accuracy and improving the working efficiencyof medical personnel.

Description

technical field [0001] The present invention relates to the technical field of deep learning, in particular to a method and system for calculating ejection fraction based on deep learning. Background technique [0002] With the development of economy and the improvement of people's living standards, the incidence of cardiovascular disease is also increasing, which is the number one threat to national health. Although the fatality rate of cardiovascular disease is high, it is also highly preventable and curable. The key is early heart disease diagnosis, screening and prediction. Ultrasound equipment has real-time requirements and has high specificity for heart diagnosis. Among the echocardiographic indicators, ejection fraction is one of the important indicators for judging the type of heart failure. Ejection fraction refers to stroke volume as a percentage of ventricular end-diastolic volume, and the normal value is 50-70%. Examination by echocardiography is one of the im...

AI Technical Summary

Problems solved by technology

However, there are the following deficiencies. First, there is a large error in the manual selection of the maximum slice position of the left ventricle (LV), which depends on experience and techniques. Second, the volume obtained by two-dimensional images is an ideal model derivation, which is consistent with the real heart. There is an error in the volume; the third is that the outline of the edge of the cavity greatly reduces the efficiency of medical staff

Images