Acute cerebral ischemia image segmentation model acquisition method and acute cerebral ischemia image segmentation method

Abstract

The invention provides an acute cerebral ischemia image segmentation model training method and an acute cerebral ischemia image segmentation method. The method comprises steps that magnetic resonance image data is acquired, and brain tissue images are extracted; normalization operation of the extracted brain tissue images is carried out; ischemic areas of the extracted brain tissue images are acquired as a positive sample, and expansion of the positive sample is carried out according to a set expansion coefficient to acquire data which is taken as a negative sample; characteristic vectors are respectively extracted from magnetic resonance images T2, DWI and ADC, the positive sample and the negative sample of asymmetric maps ASM, and normalization processing on the characteristic vectors is carried out; based on the characteristic vectors after normalization processing, an acute cerebral ischemia segmentation model is acquired through an SVM classifier. The method is advantaged in that the provided acute cerebral ischemia segmentation model can be applied to ischemic area segmentation, a segmentation speed is fast, and high accuracy is realized.

Description

technical field [0001] The invention relates to the field of medical image processing, in particular to a method for obtaining an acute cerebral ischemia image segmentation model and a method for acute cerebral ischemia image segmentation. Background technique [0002] Cerebrovascular disease is the second leading cause of death in the world. Ischemic cerebrovascular disease accounts for about 80% of all cerebrovascular diseases. According to the third national cause of death survey released by the Ministry of Health in 2008, cerebrovascular diseases have surpassed malignant tumors to become the number one cause of death in China. Acute cerebral ischemia refers to the clinical symptoms of brain dysfunction caused by insufficient blood supply to a certain part of the human brain, leading to corresponding brain function defects. Acute cerebral ischemia has an acute onset and is very harmful. Most patients will experience symptoms that affect their quality of life, such as he...

AI Technical Summary

Problems solved by technology

However, due to the rapid onset of acute cerebral ischemia, large changes in location, and difficulty in determining the shape, it is difficult to have a general segmentation method, and manual segmentation requires medical experience and anatomical knowledge, which is relatively subjective and takes time and effort.

[0005] In recent years, researchers have proposed some segmentation methods for cerebral ischemia, but there are still relatively few studies on acute cerebral ischemia within 6 hours, because compared with other general ischemia, acute cerebral ischemia has less signal in MR images. The difference is relatively large, and the boundary is blurred, so it is difficult to determine the exact position

Maier et al. proposed a method for automatic segmentation of cerebral infarction in multi-modal MR images. Using the SVM classification algorithm, the local features were extracted first. Through feature selection, the SVM algorithm was used to classify the classification model in the training stage, and the test stage will test After a series of processing, the data is input into the classification model to obtain the final classification result, but this method is not suitable for the segmentation of acute cerebral ischemia. The information displayed on the magnetic resonance image of acute cerebral ischemia is different from that of cerebral infarction.

Images