EEG Visual Classification Algorithm Based on Improved StackCNN Network and Ensemble Learning
Yang Qing1,2,3, Wang Yaqun1,2,3, Wen Dou1,2,3, Wang Ying1,2,3, Wang Xiangyu1,2,3
1. Key Laboratory of Artificial Intelligence and Smart Learning, Central China Normal University; 2. School of Computer Science, Central China Normal University; 3. National Language Resources Monitoring and Research Network Media Center, Central China Normal University
Click “Read Original” at the end to view the literature!
Table of Contents
01
Abstract
02
Chart Appreciation
03
Citation of This Article
04
Author Introduction
Abstract
Research on visual classification based on EEG signals often employs methods such as machine learning, signal processing, and statistics to analyze EEG data and extract information relevant to visual tasks. In recent years, with the rapid development of deep learning technology, deep neural networks have made significant progress in this field. Therefore, this paper addresses the issues of the limited existing research directly using image-induced EEG signals for visual classification and the low average accuracy of visual classification. A method combining Convolutional Neural Networks (CNN) and ensemble learning is designed to learn visual feature representations related to EEG signals. The model proposed in this paper can be divided into two parts: one is to decode human brain activity, which is to learn the visual classification representation of EEG signals; the second is automatic visual classification, which utilizes the learned EEG signal features for network training to achieve image classification. First, considering the issue of information loss when extracting EEG features, a K-max pooling method is incorporated into the StackCNN network to address this, combined with the Bagging algorithm to enhance the network’s generalization ability. This method is referred to as StackCNN-B, which is used to learn the visual classification representation of EEG signals. Next, to validate the performance of the StackCNN-B method in image classification, a regression based on the Deep Residual Network (ResNet) is employed for image classification. Finally, results from ablation experiments and comparisons with existing research indicate that the average accuracy of learning visual feature representations from EEG signals and the average accuracy in image classification both surpass the currently leading comparative algorithms. The research results not only confirm that EEG signals can effectively decode human brain activity related to visual recognition but also demonstrate the superiority of the proposed StackCNN-B model.
Chart Appreciation
Figure 1 Image Classification System
Figure 1 Image Classification System
Figure 2 TextCNN Network Structure
Figure 2 TextCNN Network Structure
Figure 3 StackCNN Network Structure
Figure 3 StackCNN Network
Figure 4 StackCNN-B Model Flow Chart
Figure 4 StackCNN-B Model Flow Chart
Citation of This Article
Yang Qing, Wang Yaqun, Wen Dou, et al. EEG Visual Classification Algorithm Based on Improved StackCNN Network and Ensemble Learning[J]. Journal of Zhengzhou University (Engineering Science), 2024, 45(5): 69-76.
YANG Q, WANG Y Q, WEN D, et al. EEG Visual Classification Algorithm Based on Improved StackCNN Network and Ensemble Learning[J]. Journal of Zhengzhou University(Engineering Science), 2024, 45(5): 69-76.
Author Introduction
Currently an associate professor at the School of Computer Science, Central China Normal University, and a master’s degree supervisor. Member of the Subcommittee on Literature, History, Philosophy, and Law of the National Humanities Computing Teaching Guidance Committee.
Main Research and Teaching Experience: Received a master’s degree in Computer Application Engineering from Central China Normal University in June 2000, and later obtained a Ph.D. in Management Engineering.
Research Achievements: Led projects funded by the Ministry of Education for collaborative education and laboratory projects, participated in national science and technology support projects, national social science projects, and multiple provincial and ministerial projects. Has received two second prizes and one third prize for provincial and ministerial level scientific and technological progress awards, one first prize and one third prize for teaching research achievements in Hubei Province, and was recognized as a “Good Mentor” at Central China Normal University. Holds one national invention patent and over ten copyrights for computer software from the National Copyright Administration. Has published over 40 papers, many indexed by SCI/EI, etc. Edited and contributed to over ten textbooks, and participated in drafting the “Basic Requirements for University Computer Teaching” for liberal arts majors organized by the Ministry of Education (2006, 2008, and 2010 editions). Courses taught include: Graduate courses: “Fundamentals of Computational Theory”, “Principles of Petri Nets”, “Distributed Database Technology”; Undergraduate courses: “Principles of Compilation”, “Discrete Mathematics”, “Artificial Intelligence and Social Life”, “Computer Basics”.
Recruitment Direction and Admission Requirements: Recruiting master’s students in directions such as natural language processing, smart education, data mining, etc.
Contact Information:[email protected].
Contact Information for the Journal of Zhengzhou University (Engineering Science):
Submission Website:http://gxb.zzu.edu.cn
WeChat Official Account: zdxbgxb
Contact Email: [email protected]
Contact Number: 0371-67781276
Contact Number:0371-67781277
More Exciting Content, Please Follow Us
Copyright Statement:
This article is original content from the editorial department of the Journal of Zhengzhou University (Engineering Science), and reproduction is welcome!