深度学习在神经影像中的应用研究  

Study on the Application of Deep Learning in Neuroimaging

作　　者：田苗[1] 林岚[1] 张柏雯 吴水才[1] TIAN Miao;LIN Lan;ZHANG Bai-wen;WU Shui-cai(College of Life Science and Bioengineering,Beijing University of Technology, Beijing 100124, China)

机构地区：[1]北京工业大学生命科学与生物工程学院,北京100124

出　　处：《中国医疗设备》2016年第12期4-9,共6页China Medical Devices

基　　金：国家科技支撑计划课题(2015BAI02B03);北京工业大学研究生科技基金(ykj-2016-00009);北京市自然科学基金资助项目(7143171)

摘　　要：深度学习是机器学习方法的一个重要分支,它通过基于多层神经网络的计算模型来进行学习。一个深度学习网络可以通过组合低层特征形成更抽象的高层特征,以发现数据的复杂的内在特征。由于深度卷积网络在图像处理中的出色表现,它已成为当前研究中应用最为广泛的一种深度网络。本文首先介绍了深度学习网络的结构特征和训练方法,分析了算法的优越性,之后进一步介绍了深度卷积网络,最后讨论了深度学习在神经影像领域的最新应用现状及其发展趋势。Deep learning is an important branch of a broader family of machine learning methods which attemptto learn useful features from data by using a computational model with multilayer neural networks. A deeplearning network can form more abstract high-layer features through combination of the lower layer features tofind the complex intrinsic characteristics of the data. Convolutional neural network is one of most researched deepnetworks because of its excellent performance in processing images. This paper firstly introduced the structuralcharacteristics and training methods of deep learning networks, followed by the convolutional neural network, andfinally discussed the application and development trend of deep learning network in the field of neuroimaging.

关 键 词：深度学习 卷积神经网络 神经影像 多层感知器 深度神经网络 

分 类 号：R445.2[医药卫生—影像医学与核医学]