基于深度学习的脑肿瘤患者大脑皮层表面重建  

Deep Learning-Based Cerebral Cortex Surface Reconstruction in Brain Tumor Patients

作　　者：安宁 AN Ning(Changping Laboratory,Beijing 102206,China)

机构地区：[1]昌平实验室,北京102206

出　　处：《自动化应用》2025年第2期6-9,共4页Automation Application

基　　金：昌平实验室(2021B-01-01)。

摘　　要：针对脑肿瘤患者大脑皮层表面重建难的问题,提出了一种基于深度学习技术的算法。首先,通过脑组织分割模型划分脑组织;然后,使用白质填充算法获得左右脑半球的皮层下结构掩码;再次,通过水平集回归模型预测皮层表面的水平集表示;最后,通过表面提取算法重建出大脑皮层表面。结果表明,该算法不仅解决了传统方法重建失败的问题,而且与现有深度学习方法相比,其具有更好的性能。该算法有望推动基于大脑皮层表面精准靶点定位的TMS靶向治疗方法在脑肿瘤患者中的应用。A deep learning based algorithm is proposed to address the difficulty of reconstructing the surface of the cerebral cortex in patients with brain tumors.Firstly,the brain tissue is segmented using a brain tissue segmentation model.Then,the white matter filling algorithm is used to obtain the subcortical structural masks of the left and right hemispheres of the brain.Again,predict the level set representation of the cortical surface using a level set regression model.Finally,the surface of the cerebral cortex is reconstructed using surface extraction algorithms.The results indicate that this algorithm not only solves the problem of reconstruction failure in traditional methods,but also has better performance compared to existing deep learning methods.This algorithm is expected to promote the application of TMS targeted therapy based on precise target localization on the surface of the cerebral cortex in patients with brain tumors.

关 键 词：深度学习 皮层表面重建 脑肿瘤 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]