Method, device, equipment and medium for intelligent classification of left ventricular magnetic resonance images

Abstract

The present application discloses a method, device, device, and medium for intelligent classification of left ventricular magnetic resonance images. The method includes: acquiring a first word embedding vector corresponding to the first data of a detection object, a second word embedding vector corresponding to the second data, The numerical feature vector formed by splicing the third data; extracting the short-axis video and long-axis video in the heart magnetic resonance image of the detection object, and preprocessing to obtain the target video data; splicing and inputting the above three vectors into the first feature analysis model to obtain the second The first feature analysis result and the target video data are input into the second feature analysis model to obtain the second feature analysis result, and the third feature analysis result is obtained, and input to the third feature analysis model to obtain the classification probability of the left ventricle image of the detection object value. The method uses multi-sequence nuclear magnetic imaging data and clinical data to make classification predictions. The results are more reliable and accurate, and it does not require complicated post-processing procedures, and there is no cumulative error, which improves the robustness.

Description

technical field [0001] The present disclosure relates to the technical field of image processing, in particular to the field of medical image processing, and in particular to a method, device, device and medium for intelligent classification of magnetic resonance images of the left ventricle. Background technique [0002] With the rapid development of deep learning technology, medical image processing is expected to provide low-cost, efficient and accurate auxiliary medical means. In the field of cardiac diseases, cardiac magnetic resonance has become an important standard for the classification of cardiomyopathy due to its advantages of high resolution, high soft tissue contrast and multi-sequence imaging. [0003] In related technologies, complex processes such as segmentation of MRI images and extraction of cardiac function indicators are usually required. [0004] However, in the whole process, a lot of manual operations are required, resulting in low efficiency, introd...

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned defects or deficiencies in the prior art, it is desired to provide an intelligent classification method, device, equipment and medium for magnetic resonance images of the left ventricle, which solves the problem of low efficiency due to the need for a large number of manual operations in the entire process in the related art. , introducing errors and low accuracy, making full use of multi-sequence nuclear magnetic imaging data and clinical data to make classification predictions, the results are more reliable and accurate, and by directly using end-to-end classification predictions, no complicated post-processing procedures are required. There is cumulative error, which greatly improves the robustness

Images