Classification of EEG Signals Based on Autoregressive Model and Wavelet Packet Decomposition

2018 IEEE International Conference on Pervasive Computing and Communications (PerCom)

Yao

Sheng

et al. 2018

An electroencephalography (EEG) based Brain Computer Interface (BCI) enables people to communicate with the outside world by interpreting the EEG signals of their brains to interact with devices such as wheelchairs and intelligent robots. More specifically, motor imagery EEG (MI-EEG), which reflects a subject's active intent, is attracting increasing attention for a variety of BCI applications. Accurate classification of MI-EEG signals while essential for effective operation of BCI systems, is challenging due to the significant noise inherent in the signals and the lack of informative correlation between the signals and brain activities. In this paper, we propose a novel deep neural network based learning framework that affords perceptive insights into the relationship between the MI-EEG data and brain activities. We design a joint convolutional recurrent neural network that simultaneously learns robust high-level feature presentations through low-dimensional dense embeddings from raw MI-EEG signals. We also employ an Autoencoder layer to eliminate various artifacts such as background activities. The proposed approach has been evaluated extensively on a largescale public MI-EEG dataset and a limited but easy-to-deploy dataset collected in our lab. The results show that our approach outperforms a series of baselines and the competitive state-of-theart methods, yielding a classification accuracy of 95.53%. The applicability of our proposed approach is further demonstrated with a practical BCI system for typing.

Section: Related Workmentioning

confidence: 99%

Converting Your Thoughts to Texts: Enabling Brain Typing via Deep Feature Learning of EEG Signals

2018 IEEE International Conference on Pervasive Computing and Communications (PerCom)

Yao

Sheng

et al. 2018

“…e autoregressive (AR) modeling approach, a widely used algorithm for EEG feature extraction, is also broadly combined with other feature extraction techniques to gain a be er performance [19]. For example, [29] investigated two methods EEG with AR and feature extraction combination: 1) AR model and approximate entropy, 2) AR model and wavelet packet decomposition. ey employed SVM as the classi er and showed that AR can e ectively improve classi cation performance.…”

Section: Related Workmentioning

confidence: 99%

Multi-Person Brain Activity Recognition via Comprehensive EEG Signal Analysis

Yao

Proceedings of the 14th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services

et al. 2018

An electroencephalography (EEG) based brain activity recognition is a fundamental eld of study for a number of signi cant applications such as intention prediction, appliance control, and neurological disease diagnosis in smart home and smart healthcare domains. Existing techniques mostly focus on binary brain activity recognition for a single person, which limits their deployment in wider and complex practical scenarios. erefore, multi-person and multi-class brain activity recognition has obtained popularity recently. Another challenge faced by brain activity recognition is the low recognition accuracy due to the massive noises and the low signal-to-noise ratio in EEG signals. Moreover, the feature engineering in EEG processing is time-consuming and highly relies on the expert experience. In this paper, we a empt to solve the above challenges by proposing an approach which has be er EEG interpretation ability via raw Electroencephalography (EEG) signal analysis for multi-person and multi-class brain activity recognition. Speci cally, we analyze inter-class and inter-person EEG signal characteristics, based on which to capture the discrepancy of inter-class EEG data. en, we adopt an Autoencoder layer to automatically re ne the raw EEG signals by eliminating various artifacts. We evaluate our approach on both a public and a local EEG datasets and conduct extensive experiments to explore the e ect of several factors (such as normalization methods, training data size, and Autoencoder hidden neuron size) on the recognition results. e experimental results show that our approach achieves a high accuracy comparing to competitive state-of-the-art methods, indicating its potential in promoting future research on multi-person EEG recognition.

Advances in Mechanical Engineering

“…So, its signal analysis ability is stronger. 24 The subspace U n j is defined by the closed space of the…”

Section: Wavelet Packet Decompositionmentioning

confidence: 99%

Fault diagnosis method based on wavelet packet-energy entropy and fuzzy kernel extreme learning machine

Wang

2018

Aiming at connatural limitations of extreme learning machine in practice, a new fault diagnosis method based on wavelet packet-energy entropy and fuzzy kernel extreme learning machine is proposed. On one hand, the presented method can extract the more efficient features using the wavelet packet-energy entropy method, and on the other hand, the sample fuzzy membership degree matrix U, weight matrix W which is used to describe the sample imbalance, and the kernel function are introduced to construct the fuzzy kernel extreme learning machine model with high accuracy and reliability. The experimental results of rolling bearing and check valve are obtained and analyzed in MATLAB 2010b. The results show that the proposed fuzzy kernel extreme learning machine method can obtain fairly or slightly better classification performance than the traditional extreme learning machine, kernel extreme learning machine, back propagation, support vector machine, and fuzzy support vector machine.