Fang Fang

Fang Fang

Auto-generated profile, based on publicly available information

Fang Fang

School of Mechanical Engineering and Electronic Information, China University of Geosciences, Wuhan, China

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 November 2018
Verified
Conference Paper
Closed access
JGJinliang Guo
Jinliang Guo
FFFang Fang
Fang Fang
WWWei Wang
Wei Wang
FUJI RENFUJI REN
FUJI REN
Emotion recognition is a very challenging task in the brain-computer interface field, and it is of great significance in medical, education, military, and other fields. The classification problem is the key to the field of emotion recognition research. In this paper, a classification model based on CapsNet neural network is proposed. By extracting the granger causality feature of original EEG signals, sparse group lasso algorithm is used for feature screening, and the obtained high-relevance feature subset is taken as the input of the network to achieve the final emotional classification. The experimental results show that by adjusting the model parameters and network structure, the constructed CapsNet neural network performs emotional classification on EEG signals, and obtains 88.09% and 87.37% average classification accuracy under valence and arousal emotion dimensions, compared with SVM and CNN. The classification system can obtain better results and significantly improve the performance of EEG emotional classification. This is the first time that CapsNet is applied to EEG emotional classification.

27 September 2023 in IEEE Sensors Letters
Verified
Journal Article
Closed access
YHYuxuan Hu
Yuxuan Hu
JYJun Yan
Jun Yan
FFFang Fang
Fang Fang
YWYong Wang
Yong Wang
This letter proposes a simple and effective EEG classification method using convolutional neural networks (CNN). The EEG signals are first converted to character sequences in order to simplify feature extraction and meanwhile reduce data dimensionality. The character sequences are then transformed according to binary encoding and finally fed into a specific CNN, which gives consideration to both classification accuracy and computation efficiency. Experimental simulations show the proposed method has the advantages of simple preprocessing and lightweight network structure, the average accuracy on two BCI Competition datasets can reach 94.36% and 81.8%, while the average accuracy on the self-made dataset is 71.67%, outperforming traditional methods. Furthermore, the proposed algorithm has been successfully applied to control the robotic arm of an EEG-based manipulator system, further validating its feasibility.