【计算机科学】【2017.08】脑电信号分类中的深度学习与迁移学习

本文为美国内布拉斯加大学（作者：Jacob M. Williams）的硕士论文，共82页。

尽管在其他多种空间和时间序列数据中实现了最先进的分类精度，但深度学习很少用于脑电图（EEG）信号的分类。相反，大多数研究继续使用手工特征提取，然后使用传统的分类器，如SVMs或逻辑回归。这主要是由于每个实验的样本数量少、数据的高维性质以及难以找到适合于脑电图信号分类的深度学习架构。本文将几种深度学习架构与传统的视觉诱发脑电信号分类方法进行了比较。我们发现使用长短期记忆单元（LSTM）的深度学习体系架构优于传统方法，而小型卷积体系架构的性能与传统方法相当。我们还探讨了通过跨多个科目的训练和对特定科目的改进来使用迁移学习，这种形式的迁移学习进一步提高了深度学习模型的分类精度。

Deep learning is seldom used in the classification of electroencephalography (EEG) signals, despite achieving state of the art classification accuracies in other spatial and time series data. Instead, most research has continued to use manual feature extraction followed by a traditional classifier, such as SVMs or logistic regression. This is largely due to the low number of samples per experiment, high-dimensional nature of the data, and the difficulty in finding appropriate deep learning architectures for classification of EEG signals. In this thesis, several deep learning architectures are compared to traditional techniques for the classification of visually evoked EEG signals. We found that deep learning architectures using long short-term memory units (LSTMs) outperform traditional methods, while small convolutional architectures performed comparably to traditional methods. We also explored the use of transfer learning by training across multiple subjects and refining on a particular subject. This form of transfer learning further improved the classification accuracy of the deep learning models.

1 引言

2 相关工作与研究背景

3 问题定义与使用方法

4 数据集、具体实现与结果

5 总结与未来工作展望

下载英文原文地址：

http://page2.dfpan.com/fs/7lcje221f291d650b72/

更多精彩文章请关注微信号：