Method for predicting healthy radius of blood vessel path, method for predicting candidate stenosis of blood vessel path, and blood vessel stenosis prediction device

Abstract

The present disclosure provides a method for predicting the healthy radius of a blood vessel path, a method for predicting the candidate stenosis of the blood vessel path, and a blood vessel stenosisprediction device. The method for predicing the health radius includes steps of: acquiring the blood vessel radius of the blood vessel path; detecting a peak value of the blood vessel radius of the blood vessel path by using a processor; and predicting the healthy radius by performing regression on the radius peak value by using the processor. The blood vessel stenosis prediction device includes aprocessor configured to execute an executable instruction to: acquire an image unit along a centerline within a candidate stenosis range of each blood vessel path, and determine, by using a learningnetwork based on the acquired image unit, the stenosis of each blood vessel path, wherein the learning network formed based on a convolutional neural network and a recurrent neural network. The devicecan automatically determine the candidate stenosis, detect the stenosis of the candidate stenosis range, can significantly reduce the calculation load, improves the detection efficiency and effectively avoids missed detection.

Description

technical field [0001] The present disclosure relates to a method and device for predicting vascular physiological parameters, and more specifically, to a method for predicting the healthy radius of a vascular path (a section of blood vessel), a method for predicting a candidate stenosis of a vascular path, and using machine learning to A device for predicting the degree of narrowing of blood vessels. Background technique [0002] Recent studies have shown that the stenosis of blood vessels is of great significance in the diagnosis and prevention of cardiovascular diseases. Usually the stenosis of blood vessels is manually determined, that is, experts judge by manual observation of coronary CT angiography (CTA) and / or quantitative coronary angiography (QCA), but this judgment is time-consuming, tedious and wastes a lot of time. resources and relies heavily on the experience of experts. [0003] Although some automatic or semi-automatic methods for determining the degree of...

AI Technical Summary

Problems solved by technology

The independent training of the modules requires corresponding training data, and the requirements for the training data are high; and the training cannot learn from each other, and the optimal performance of the system as a whole cannot be achieved.

Images