Method and system for automatically segmenting blood vessel in medical image by using machine learning and image processing algorithm
A method for automatically segmenting three-dimensional blood vessel data from three-dimensional medical image data of a patient through the use of a computer is provided. The method includes: receiving the three-dimensional medical image data of the patient; generating three-dimensional shape machine-learning blood vessel data from the received three-dimensional medical image data through the use of a machine-learned segmentation program so as to generate three-dimensional blood vessel data; and generating corrected three-dimensional shape blood vessel data from the received three-dimensional medical image data and the generated three-dimensional shape machine-learning blood vessel data through the use of an image processing program.
1. A method for automatically segmenting three-dimensional blood vessel data from three-dimensional medical image data of a patient through the use of a computer, comprising:
receiving the three-dimensional medical image data of the patient;
generating three-dimensional shape machine-learning blood vessel data from the received three-dimensional medical image data through the use of a machine-learned segmentation program so as to generate three-dimensional blood vessel data; and
generating corrected three-dimensional shape blood vessel data from the received three-dimensional medical image data and the generated three-dimensional shape machine-learning blood vessel data through the use of an image processing program,
wherein the three-dimensional machine-learning blood vessel data is composed of at least one noise data set other than a blood vessel region and a data set in which the blood vessel region is missing, and
the image processing program is configured to compare the received three-dimensional medical image data and the generated three-dimensional shape machine-learning blood vessel data to match the blood vessel region, supplement the missing data set, and remove the noise data set, so as to generate corrected three-dimensional shape blood vessel data.
2. The method of claim 1 , wherein the machine-learned segmentation program includes a FCN algorithm.
3. The method of claim 2 , wherein the FCN algorithm of the machine-learned segmentation program is configured to convert axial slice images of three-dimensional medical images into bitmap images, label a blood vessel region of each of the converted bitmap images, convert each of the labeled bitmap images into a mask image to utilize the mask image as learning data, and perform learning by using each of the labeled bitmap images and the mask image as a pair.
4. The method of claim 3 , wherein the machine-learned segmentation program further includes a GAN algorithm.
5. The method of claim 1 , wherein the image processing program is configured to supplement the missing data set through the use of a region growing algorithm for the blood vessel region of the received three-dimensional medical image data by using coordinate information of the generated three-dimensional shape blood vessel data as a seed, calculate a volume of the three-dimensional shape blood vessel data sets to which the missing data region is connected, determine blood vessel data whose calculated volume is equal to or less than a predetermined value as a noise data set, and remove the noise data set.
6. A system for automatically segmenting three-dimensional blood vessel data from three-dimensional medical image data of a patient through the use of a computer, comprising:
a computer installed with a computer program for performing the method of claim 1 .