Automatic quantitative analysis method of blood flow in intracranial and extracranial vascular anastomosis

Abstract

The present invention provides a method for automatic quantitative analysis of blood flow in intracranial and extracranial vascular anastomosis, comprising the steps of: S1, extracting a frame of video images from the preoperative and postoperative ICG-VA videos of several patients; S2, training a multi-task Unet network, It is used to segment all blood vessels and recipient blood vessels in a frame of ICG-VA video image; S3, based on the trained network, segment all blood vessels and recipient blood vessels of a frame of video images of patients before and after surgery; S4, based on sift Algorithm, registering the preoperative and postoperative video images of the patient; S5, analyzing the blood flow direction of the recipient vessel before and after the patient based on the HS optical flow method. The invention can not only analyze the blood perfusion of recipient blood vessels, but also automatically detect and match the blood vessels in the images before and after surgery, measure the blood flow direction and intensity curve of any blood vessel, and provide the corrected blood perfusion change diagram before and after surgery , enabling the operator to visually assess changes in blood perfusion caused by the anastomosis.

Description

technical field [0001] The invention relates to the technical field of computer-aided diagnosis, in particular to an automatic quantitative analysis method for blood flow in intracranial and extracranial vascular anastomosis. Background technique [0002] Superficial temporal artery-middle cerebral artery anastomosis (STA-MCA) was first reported by Spetzler and Martin in 1986, and has been widely used in blood flow reconstruction of clinical hemorrhage or ischemic cerebrovascular disease and complex intracranial aneurysm. Raabe et al. demonstrated for the first time that indocyanine green (ICG) dynamically flows through anastomotic vessels, which can assist in determining the patency and direction of blood flow during surgery. A large number of studies have demonstrated that intraoperative indocyanine green angiography (ICG-VA) and existing commercial software can be used to generate time-lapse color maps to obtain regional cerebral blood flow (CBF) and evaluate postoperativ...

AI Technical Summary

Problems solved by technology

[0003] However, the commercial software used as image post-processing cannot realize continuous real-time and dynamic analysis of CBF, and cannot automatically detect the blood flow direction of blood vessels before and after surgery.

[0004] Because there are many blood vessels with large diameter differences and low resolution in ICG-VA images, the existing commercial software has a low recognition rate of blood vessels and the recognition accuracy is not high

Images