Article
Computer Science, Artificial Intelligence
Zumray Dokur, Tamer Olmez
Summary: It is found that using an augmentation process can achieve high success rates on small datasets similar to common spatial patterns (CSP) for classification of MI EEG signals, instead of just increasing classification performance through transformations. Also, by modifying the DNN structure and using a minimum distance network following the CNN, the classification performance can be further increased.
APPLIED SOFT COMPUTING
(2021)
Article
Computer Science, Artificial Intelligence
Sheng-wei Fei, You-bing Chu
Summary: This study proposes a classification method based on WT-PSR-SVD for multi-layer twin support vector machines, which can control the actions of devices more accurately. Experimental results demonstrate that the proposed method has better classification ability for motor imagery EEG signals compared to traditional methods.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Sujin Bak, Jichai Jeong
Summary: Human brain activities, specifically electroencephalogram (EEG) signals, are being explored as a secure biometric approach for user identification due to their sensitivity, secrecy, and difficulty to replicate. This study proposes an EEG-MI methodology that utilizes optimized feature extraction methods and classifiers to improve user-aware accuracy. The results show that the proposed methodology achieved high user identification accuracies using the support vector machine (SVM) and Gaussian Naive Bayes (GNB) classifiers.
Review
Computer Science, Artificial Intelligence
Hamdi Altaheri, Ghulam Muhammad, Mansour Alsulaiman, Syed Umar Amin, Ghadir Ali Altuwaijri, Wadood Abdul, Mohamed A. Bencherif, Mohammed Faisal
Summary: This paper systematically reviews the DL-based research for MI-EEG classification in the past ten years. It analyzes and discusses DL techniques applied in MI classification from various perspectives and addresses key questions in DL-based MI classification. The paper also summarizes MI-EEG applications and explores public MI-EEG datasets.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Chun-Na Li, Yuan-Hai Shao, Huajun Wang, Yu-Ting Zhao, Naihua Xiu, Nai-Yang Deng
Summary: This paper investigates the general forms and characteristics of nonparallel support vector machines (NSVMs) and categorizes them into two types. It reveals the advantages and defects of different types and points out the inconsistency problems. Based on this observation, a novel max-min distance-based NSVM is proposed with desired consistency. The proposed NSVM has the consistency of training and test and the consistency of metric, and it assigns each sample an ascertained loss.
APPLIED SOFT COMPUTING
(2023)
Article
Engineering, Biomedical
Kostas Georgiadis, Dimitrios A. Adamos, Spiros Nikolopoulos, Nikos Laskaris, Ioannis Kompatsiaris
Summary: In this work, the Graph Slepian functions technique is utilized for robust decoding of motor imagery brain activity, incorporating domain knowledge and providing flexibility in data analysis. This algorithmic pipeline enhances information conveyed by multichannel signals and relates to participant's intention, ultimately improving MI decoding efficacy. The proposed technique is computationally efficient and outperforms popular alternatives in the field, making it suitable for real-time implementations.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2021)
Article
Computer Science, Interdisciplinary Applications
Maliheh Abbaszadeh, Saeed Soltani-Mohammadi, Ali Najah Ahmed
Summary: This article introduces the application of the support vector classifier in geological modeling and proposes an improved method based on particle swarm optimization to select the best model parameters. Through the application in the modeling process of the Iju porphyry copper deposit, the effectiveness and superiority of this method are demonstrated.
COMPUTERS & GEOSCIENCES
(2022)
Article
Computer Science, Information Systems
Md Khademul Islam Molla, Sanjoy Kumar Saha, Sabina Yasmin, Md Rabiul Islam, Jungpil Shin
Summary: This study utilizes electroencephalography (EEG) to classify motor imagery tasks, extracting features and selecting classifiers using discrete wavelet transform and common spatial pattern. Experimental results show that the proposed method outperforms other algorithms.
Article
Biochemistry & Molecular Biology
Hongli Li, Wei Guo, Ronghua Zhang, Chunbo Xiu
Summary: This article introduces a novel algorithm based on VLPSO and MFDF, which can improve the classification accuracy of motor imagery EEG signals by fusing multi-domain features and correcting prediction differences.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
(2021)
Article
Mathematics, Interdisciplinary Applications
Eda Dagdevir, Mahmut Tokmakci
Summary: A novel methodology based on Truncation Thresholds method, Empirical Mode Decomposition, and statistical Common Spatial Pattern feature extraction is proposed for EEG signals classification. The proposed method achieved high accuracy in distinguishing left and right hand imaginary movements from EEG signals.
CHAOS SOLITONS & FRACTALS
(2021)
Article
Engineering, Electrical & Electronic
Seffi Cohen, Or Katz, Dan Presil, Ofir Arbili, Lior Rokach
Summary: Excessive drinking is a significant risk factor for various health complications. The current diagnostic methods for alcoholism, which rely on blood tests and subjective questionnaires, have limitations. This study aims to develop new EEG classification methods to improve accuracy in predicting alcoholism. By converting temporal data into images and utilizing ensemble classification models, our algorithm achieves a cross-validation classification accuracy of 85.52%, outperforming the state-of-the-art method by 4.33%.
IEEE SENSORS JOURNAL
(2023)
Article
Engineering, Biomedical
Hongli Li, Man Ding, Ronghua Zhang, Chunbo Xiu
Summary: A neural network feature fusion algorithm combining CNN and LSTM has been proposed to improve the accuracy of motor imagery EEG classification, providing new ideas for related research. The average accuracy and Kappa value were found to be 87.68% and 0.8245, respectively.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2022)
Article
Computer Science, Artificial Intelligence
Rongda Chen, Zhixia Yang, Junyou Ye
Summary: This article discusses the challenges of using support vector machine (SVM) models in multiview learning and proposes two multiview classifiers, C-MKNSVM and ?-MKNSVM, which overcome the difficulties by using kernel-free techniques. Experimental results show that these classifiers outperform traditional MVL classifiers.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2023)
Article
Engineering, Marine
Liadira Kusuma Widya, Chang-Hwan Kim, Jong-Dae Do, Sung-Jae Park, Bong-Chan Kim, Chang-Wook Lee
Summary: The study utilized support vector machine (SVM) technologies with corrected satellite imagery data to accurately identify the distribution of seagrasses.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2023)
Correction
Computer Science, Artificial Intelligence
Ala' M. Al-Zoubi, Mohammad A. Hassonah, Ali Asghar Heidari, Hossam Faris, Majdi Mafarja, Ibrahim Aljarah
Summary: The affiliation information of authors Ali Asghar Heidari and Majdi Mafarja was mistakenly published during typesetting and has been corrected.
Review
Computer Science, Artificial Intelligence
Yukai Zhou, Qingshan She, Yuliang Ma, Wanzeng Kong, Yingchun Zhang
Summary: This paper introduces a novel semi-supervised broad learning system TSS-BLS, which obtains pseudo-labels of unlabeled samples using joint distribution adaptation algorithm and constructs a TSS-BLS system containing both labeled and pseudo-label information. Experimental results show that TSS-BLS outperforms BLS and GSS-BLS on average, providing a safe and efficient method for EEG classification.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Biochemical Research Methods
Ming Meng, Xu Yin, Qingshan She, Yunyuan Gao, Wanzeng Kong, Zhizeng Luo
Summary: The paper introduces a TDDO-SRC method for motor imagery EEG pattern recognition, showing its superior performance and comparison results. The quality of dictionary construction has a significant impact on the robustness of SRC, and TDDO-SRC significantly improves classification accuracy compared to the original SRC.
JOURNAL OF NEUROSCIENCE METHODS
(2021)
Article
Engineering, Electrical & Electronic
Chujie Wei, Tao Fang, Yingle Fan, Wei Wu, Ming Meng, Qingshan She
Summary: This study proposes a new method of image contour detection based on binocular parallax, which can extract the primary contour of an image and suppress local texture through innovative adjustments of opponent cell connection weights, binocular parallax energy model, and multi-scale receptive field fusion strategy. It provides a new idea for subsequent studies on the higher visual cortex's image understanding and visual cognition.
SIGNAL IMAGE AND VIDEO PROCESSING
(2022)
Article
Biochemical Research Methods
Yinhao Cai, Qingshan She, Jiyue Ji, Yuliang Ma, Jianhai Zhang, Yingchun Zhang
Summary: Brain computer interface (BCI) is a technology that utilizes brain signals for human-computer interaction. However, building a generic EEG recognition model is challenging due to non-stationarity, subject variations, and extensive training requirements. To address this, a manifold embedded transfer learning (METL) framework is proposed to transfer calibration information from existing subjects to new subjects. Experimental results confirm the effectiveness of METL, especially when there are limited labeled samples in the target domain. METL outperforms conventional methods in terms of classification accuracy.
JOURNAL OF NEUROSCIENCE METHODS
(2022)
Article
Computer Science, Artificial Intelligence
Bi-Peng Chen, Yun Chen, Guo-Qiang Zeng, Qingshan She
Summary: This article focuses on the intelligent optimization issue using PEO-FOCNN, which combines fractional order convolutional neural networks (FOCNNs) with population extremal optimization (PEO). The Caputo fractional-order gradient method (CFOGM) is introduced to improve the optimization performance of FOCNN. The experiments demonstrate the superiority of PEO-FOCNN over other optimization algorithms on the MNIST dataset.
Article
Engineering, Electrical & Electronic
Ming Meng, Zhichao Dong, Yunyuan Gao, Qingshan She
Summary: This study proposes an optimal channel and frequency band-based CSP feature selection method for improving the performance of EEG feature extraction in brain-computer interface based on motor imagery. The method selects the optimal channels using correlation coefficient and chooses subbands with higher power spectrum density for CSP feature extraction. Fisher ratio is utilized for further feature selection. Experimental results on BCI competition datasets demonstrate the rationality and effectiveness of the proposed method.
INTERNATIONAL JOURNAL OF IMAGING SYSTEMS AND TECHNOLOGY
(2023)
Article
Engineering, Biomedical
Lei Chen, Qingshan She, Ming Meng, Qizhong Zhang, Jianhai Zhang
Summary: This paper proposes a novel method of similarity constraint style transfer mapping (SCSTM) and domain selection strategy with geodesic flow kernel (DSSWGFK) to address the problem of limited labeled data in emotion recognition based on EEG. Experimental results show that the proposed method achieves better average classification accuracy compared to existing methods.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Engineering, Biomedical
Qingshan She, Tie Chen, Feng Fang, Jianhai Zhang, Yunyuan Gao, Yingchun Zhang
Summary: This study proposes an improved domain adaption network based on Wasserstein distance to enhance the performance of Motor Imagery (MI) classification on a single subject using EEG data.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2023)
Article
Engineering, Biomedical
Yuliang Ma, Weicheng Zhao, Ming Meng, Qizhong Zhang, Qingshan She, Jianhai Zhang
Summary: This paper proposes a novel method to address the problem of the decline in accuracy of cross-subject emotion recognition due to negative data transfer in the source domain. The method, named cross-subject source domain selection (CSDS), consists of three parts: establishing a Frank-copula model to study the correlation between the source and target domains, improving the calculation method for Maximum Mean Discrepancy to determine distances between classes in a single source, and using Local Tangent Space Alignment to provide a low-dimensional linear estimation of the local geometry of nonlinear manifolds during transfer learning.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Qingshan She, Yinhao Cai, Shengzhi Du, Yun Chen
Summary: This paper proposes a multi-source manifold feature transfer learning (MMFT) framework for transferring multi-source knowledge in EEG signal classification. Experimental results show that MMFT outperforms other methods in terms of classification accuracy and computational efficiency.
Article
Computer Science, Information Systems
Yunyuan Gao, Yici Liu, Qingshan She, Jianhai Zhang
Summary: Transfer learning in brain-computer interfaces (BCIs) has potential applications. A multi-manifold embedding domain adaptive algorithm is proposed to effectively transfer data from a source to target domain. The algorithm aligns EEG covariance matrix in Riemannian manifold, extracts characteristics in tangent space, and maps them to Grassmann manifold for a common feature representation. Geometric and statistical attributes of EEG data are considered in domain adaptation, resulting in satisfactory results on EEG transfer learning.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Engineering, Electrical & Electronic
Qingshan She, Chenqi Zhang, Feng Fang, Yuliang Ma, Yingchun Zhang
Summary: This article proposes a new emotion recognition method based on a multisource associate domain adaptation (DA) network, which solves the issue of variability in emotion recognition models across different subjects and sessions in brain-computer interface systems. The method constructs separate branches for multiple source domains and extracts domain-specific features to achieve emotion recognition through weighted scoring and multiple source classifiers. Experimental results show that the proposed method outperforms state-of-the-art DA methods in classification.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Mathematical & Computational Biology
Shaojun Zhu, Jinhui Zhao, Yating Wu, Qingshan She
Summary: This paper proposes a new and effective method for entropy transfer by applying R-vine copula function estimation. The experiment results show that the proposed method can accurately infer complex causal coupling and provide a new theoretical perspective for the diagnosis of neuromuscular fatigue and sports rehabilitation.
MATHEMATICAL BIOSCIENCES AND ENGINEERING
(2022)
Article
Mathematical & Computational Biology
Xu Yin, Ming Meng, Qingshan She, Yunyuan Gao, Zhizeng Luo
Summary: A new optimal channel-based sparse time-frequency blocks common spatial pattern (OCSB-CSP) feature extraction method is proposed to improve the classification accuracy and computational efficiency of the model. Comparative experiments on two public EEG datasets show that the proposed method outperforms other winner methods in terms of classification performance. This provides a new idea for enhancing BCI applications.
MATHEMATICAL BIOSCIENCES AND ENGINEERING
(2021)
Article
Mathematical & Computational Biology
Ming Meng, Luyang Dai, Qingshan She, Yuliang Ma, Wanzeng Kong
Summary: A novel crossing time windows optimization method for mental arithmetic (MA) based BCI is proposed, combining EEG and fNIRS signals separately and extracting features using FLFS and LDA. The classification accuracy of the proposed method on the MA dataset is 92.52 +/- 5.38%, demonstrating the rationality and superiority of this approach.
MATHEMATICAL BIOSCIENCES AND ENGINEERING
(2021)
