Brain tumor image segmentation method, device and equipment based on deep learning, and medium

Abstract

The method is applied to the technical field of artificial intelligence, relates to the technical field of blockchain, and discloses a brain tumor image segmentation method, device and equipment basedon deep learning, and a medium. The method includes, in part, the following steps: obtaining a multi-mode brain nuclear magnetic resonance image; preprocessing the brain nuclear magnetic resonance image, so as to obtain a target image with the skull part being removed; and finally, inputting the target image into a preset brain tumor segmentation model, so as to obtain a brain tumor image segmentation result, wherein the preset brain glioma segmentation model is a deep learning model obtained by performing cross validation training according to an adaptive segmentation framework and a brain nuclear magnetic resonance image without a skull part, and the adaptive segmentation framework comprises a plurality of different types of U-Net models and U-Net integrated models. According to the invention, the optimal network structure for prediction in the plurality of models can be automatically selected according to the result of cross validation, and the segmentation performance of the preset brain tumor segmentation model is improved, so the accuracy of brain tumor image segmentation is improved.

Description

technical field [0001] The present invention relates to the technical field of brain tumor image segmentation, in particular to a method, device, equipment and medium for brain tumor image segmentation based on deep learning. Background technique [0002] Glioma is the most common primary intracranial malignant tumor. According to the degree of malignancy, it can be divided into WHO grades I-IV. As the grade increases, the degree of malignancy increases. Treatment methods for gliomas of different grades and gene mutations difference in prognosis. Therefore, if the tumor area and tumor grade can be accurately segmented and judged before surgical treatment, it will help guide the treatment plan and the selection of surgical resection area, which is of great value in improving the treatment effect and prognosis of patients. [0003] In the traditional glioma segmentation process, clinicians need to use image processing software to manually or semi-automatically segment the pat...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a brain tumor image segmentation method, device, equipment and medium based on deep learning to solve the problem that in the existing brain tumor image segmentation method, it is difficult to find the optimal network structure during the training process of the segmentation model, resulting in inaccurate model accuracy. High, which leads to the problem of low segmentation accuracy of brain tumor images

Images