Tumor image brain region segmentation method and system based on image completion

Abstract

The invention discloses a tumor image brain region segmentation method and system based on image completion. The method comprises the steps of obtaining a to-be-segmented tumor image and a tumor mask;inputting the to-be-segmented tumor image and the tumor mask into a completion network to obtain a completed to-be-segmented image; wherein the completion network is obtained by training a PconvUnetnetwork by taking a normal brain image and a random mask as a training set and taking the minimum overall loss function as a target; inputting the completed to-be-segmented image into the segmentationnetwork to obtain brain partitions of the to-be-segmented tumor image; wherein the segmentation network is obtained by training the Unet network by taking a normal brain image and a corresponding image label as a training set and taking the minimum similarity measure loss function or the minimum cross entropy loss function as a target. The accuracy of brain region segmentation of the tumor imagecan be improved.

Description

technical field [0001] The present invention relates to the field of image segmentation, in particular to a method and system for brain region segmentation of tumor images based on image complementation. Background technique [0002] Brain tumors mostly grow in the cranial cavity, also known as intracranial tumors and brain cancers, which can originate from the brain, meninges, nerves, blood vessels, and brain appendages, or invade the brain from other tissues or organs of the body. The incidence of brain tumors is about 1.9 to 5.4 people / (year·100,000 people), accounting for 1% to 3% of all kinds of tumors in the whole body. Clinically, brain imaging data obtained by CT (Computed Tomography) and MRI (Magnetic Resonance Imaging) are mainly used for medical image processing. [0003] Due to the influence of image acquisition equipment, medical images (such as MRI images) will have problems such as low contrast, low signal-to-noise ratio, and low light intensity; and the orga...

AI Technical Summary

Problems solved by technology

[0005] The existing brain region segmentation methods for tumor images have the following disadvantages: 1) Manually label tumor brain regions: manual labeling of tumor brain regions faces the problems of high labor costs and strong subjectivity

2) Segmentation through the traditional brain partition model: this method has a poor segmentation result for the tumor part due to the large gap between the tumor image and the normal brain image in the tumor part

3) Segmentation by registration: The registration of brain tumor images has a large gap with normal brain images due to the existence of tumor parts, and the registration effect is poor, which in turn affects the segmentation effect

4) Training tumor image brain segmentation network: Due to the variety of tumor locations and shapes and the lack of tumor data, this method makes network learning tumor image brain segmentation difficult.

Images