Vascular model extraction method based on depth neural network

Abstract

The invention discloses a blood vessel model extraction method based on a depth neural network. The method comprises the following steps: 1. Blood vessel data is enhanced by T_Frangi algorithm according to the spatial scale theory so as to enhance the blood vessel data; 2, reserve candidate data; 3, calculate that characteristics of the blood vessel connection region; Step 4: Deep neural network training, using the tuple of blood vessel features as input to train the neural network, so as to obtain the blood vessel extraction model. The blood vessel model extraction method based on the depth neural network, In the candidate region of blood vessel after depth enhancement, Using the connectivity of blood vessels, the five-element feature set of each connected region is constructed, and the neural network model is trained to extract blood vessels. This method does not need to train the whole volume data, only needs to consider the calculation of candidate blood vessel connectivity regions. It can effectively remove the outliers, and has high extraction accuracy and flexibility.

Description

Technical field [0001] The invention relates to the technical field of medical image processing, in particular to a method for extracting a blood vessel model based on a deep neural network. Background technique [0002] Vascular extraction is the cornerstone of many medical imaging applications, such as acute ischemic stroke. Vascular extraction is very important for the quantification of vascular occlusion and the evaluation of collateral flow. In the diagnosis of coronary atherosclerosis, vascular extraction is the detection of luminal stenosis. An essential step. [0003] Vessel extraction algorithms can generally be divided into two categories: one is active contour models, such as geodesic active contour models. Because the gray value at the edge of the blood vessel is similar to the gray value of the surrounding tissue, it is difficult to accurately segment the blood vessel by using the boundary gradient and regional information; and the active contour model needs to be adj...

AI Technical Summary

Problems solved by technology

[0005] It is difficult to take into account the three-dimensional characteristics of blood vessels in general two-dimensional extraction, and the operation efficiency of three-dimensional extraction is not high, and there are different degrees of isolated points or non-vascular areas in the extraction results

Images