Intravascular plaque attribute analysis method based on depth migration learning

Abstract

The invention discloses an intravascular plaque attribute analysis method based on depth migration learning, belonging to the medical science image processing field. The invention particularly relatesto an intravascular plaque attribute analysis method based on depth migration learning. The intravascular plaque attribute analysis method comprises steps of obtaining multi-mode intravascular imagedata in a clinical mode, artificially marking an attribute of an intravascular plaque, performing pre-processing on a marked intravascular image, using intravascular image data which is pre-processedas an input of a depth convolution neural network, adopting a learning method with supervision to perform depth convolution neural network training, adopting a back-propagation-algorithm-based stochastic gradient descent mode to perform network parameter learning, adopting trained multiple cross-modal prediction models to perform voting fusion prediction on the intravascular image obtained from the step 1 and generating a plaque kind probability graph. The intravascular plaque attribute analysis method based on depth migration learning solves a problem that the prior art is labor-consuming andlow in speed and has artificial difference. The intravascular plaque attribute analysis method based on depth migration learning can be used for intravascular image processing.

Description

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to view more

AI Technical Summary

Benefits of technology

This patented technique allows researchers to analyze blood vessel images from various medical devices without having them physically touch their body or rely heavily on subjective judgment. It uses advanced techniques like computerized tomographic scans and magnetic resonance imagings to enhance visualization capabilities while reducing errors associated therewith. By doing these processing steps beforehand, we are able to identify specific areas containing abnormal tissue patterns which could indicate potential future complications during treatment procedures. Overall, our technical effect lies in providing faster and easier ways to diagnose disease conditions earlier by gathering relevant patient information about those who have had previous treatments.

Problems solved by technology

This patented technical solution describes how current methods involve manually extracting characteristic values from images obtained during examinations like angiography, CT scans, magnetic resonance imaging (MRI), ultrasound, X-ray fluorescence testing, positron emission tomography (PET) scanings, and biomedical techniques can help identify areas where abnormalities may be present early enough to take appropriate actions based upon these findings. However, this method requires significant human effort due to its slowness and potential variations between individuals who make up their own unique dataset.

Images