深度学习在脑胶质瘤影像学的研究进展

Advances in deep learning for brain glioma imaging

深度学习是基于多层神经网络计算模型发现数据内复杂特征的一种深度网络,较多应用于医学图像的分割与分类中,在各类脑胶质瘤的研究中也有许多成果。本文就深度学习在脑胶质瘤的准确分割定位、组织遗传学特征预测及预后评估等方面展开综述,总结深度学习在脑胶质瘤影像图像分割与分类的研究进展,从而为胶质瘤患者的精准诊断、个体化治疗提供新思路。

Deep learning is a kind of deep network that can discover the inherent complex features of data based on a multilayer neural network computing model.It is mostly used in the segmentation and classification of medical images,and it also makes lots of achievements in the research of various types of gliomas.Herein the applications of deep learning in the accurate segmentation and positioning,prediction of tissue genetic features,and prognostic evaluation of brain gliomas are reviewed,and the recent advances in deep learning for image segmentation and classification of gliomas are summarized,so as to provide new ideas for the accurate diagnosis and individualized treatment of glioma patients.