Article
Computer Science, Information Systems
Mohammad Asif, Sudhakar Mishra, Majithia Tejas Vinodbhai, Uma Shanker Tiwary
Summary: Emotion recognition using EEG signals is a promising field in Brain-Computer Interfaces. To overcome the limitations of existing emotion databases, we designed an experiment where participants freely reported their emotional feelings while watching emotional stimuli. Our dataset, DENS, showed higher accuracy in classifying emotional events compared to benchmark datasets DEAP and SEED.
Article
Engineering, Multidisciplinary
Raja Majid Mehmood, Muhammad Bilal, S. Vimal, Seong-Whan Lee
Summary: By using a novel Hjorth-feature-based emotion recognition model, this study explores a wider set of emotion classes and achieves better accuracy in EEG-based emotion recognition.
Article
Neurosciences
Kai Yang, Li Tong, Ying Zeng, Runnan Lu, Rongkai Zhang, Yuanlong Gao, Bin Yan
Summary: Recent studies have shown that recognizing and monitoring different valence emotions can effectively prevent human errors caused by cognitive decline. This study explores effective electroencephalography (EEG) features for recognizing different valence emotions. The results show that first-order difference, second-order difference, high-frequency power, and high-frequency differential entropy features perform better in emotion recognition. Time-domain features, especially first-order difference and second-order difference features, have shorter computing time, making them suitable for real-time emotion recognition applications. Features extracted from the frontal, temporal, and occipital lobes are more effective in recognizing different valence emotions. Furthermore, when the number of electrodes is reduced by 3/4, using features from 16 electrodes located in these brain regions achieves a classification accuracy of 91.8%, only about 2% lower than using all electrodes. These findings provide important guidance for feature extraction and selection in EEG-based emotion recognition.
FRONTIERS IN NEUROSCIENCE
(2022)
Review
Mathematical & Computational Biology
Haoran Liu, Ying Zhang, Yujun Li, Xiangyi Kong
Summary: This paper discusses the common steps and existing methods of EEG-based emotion recognition algorithms, providing foundational theory and references for future research. Emotion is closely related to safety psychology.
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE
(2021)
Article
Multidisciplinary Sciences
S. Mazzacane, M. Coccagna, F. Manzella, G. Pagliarini, V. A. Sironi, A. Gatti, E. Caselli, G. Sciavicco
Summary: The study focuses on analyzing electroencephalogram signals recorded from participants during experiences to understand the brain processes behind their physical and emotional engagement. The researchers use a novel symbolic machine learning technique to extract information from unstructured data and express it as logical rules, relating voltage patterns at specific frequencies and electrodes to liking or disliking experiences. This study contributes to the field of neuroaesthetics, providing insights into the brain mechanisms associated with aesthetic perceptions in human subjects.
Article
Chemistry, Analytical
Sungkyu Kim, Tae-Seong Kim, Won Hee Lee
Summary: Deep learning-based emotion recognition using EEG has gained attention. This study proposes a novel 3D convolutional neural network model for EEG-based emotion recognition, which achieves high accuracy and reduces computational complexity.
Article
Computer Science, Artificial Intelligence
Dongmin Huang, Sentao Chen, Cheng Liu, Lin Zheng, Zhihang Tian, Dazhi Jiang
Summary: Neuroscience research has found that the left and right hemispheres of the human brain respond differently to emotions, which is crucial for emotion recognition. The proposed BiDCNN model effectively learns these differences and achieves state-of-the-art performance in emotion recognition tasks. The model demonstrates high accuracy rates in both subject-dependent and subject-independent experiments.
Article
Computer Science, Theory & Methods
Yishu Liu, Guifang Fu
Summary: This paper introduces a method for human emotion recognition using multi-channel features learned from EEG signal and textual features, improving emotion classification accuracy by fusing different statistical features in the time domain.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2021)
Article
Chemistry, Analytical
Rajamanickam Yuvaraj, Prasanth Thagavel, John Thomas, Jack Fogarty, Farhan Ali
Summary: Advances in signal processing and machine learning have accelerated EEG-based emotion recognition research. This study compared the classification accuracy of various sets of EEG features to identify emotional states. By evaluating the performance on five independent datasets, it was found that the FD-CART feature-classification method achieved the highest accuracy for valence and arousal. These findings suggest the reliability of the FD features derived from EEG data for emotion recognition, and may contribute to the development of a real-time EEG-based emotion recognition system.
Article
Computer Science, Artificial Intelligence
Rui Li, Chao Ren, Yiqing Ge, Qiqi Zhao, Yikun Yang, Yuhan Shi, Xiaowei Zhang, Bin Hu
Summary: This study proposed a novel emotion recognition model called MTLFuseNet, which uses deep latent feature fusion of EEG signals and multi-task learning. Through unsupervised learning by a variational autoencoder (VAE), MTLFuseNet learned spatio-temporal latent features of EEG, and through supervised learning by a graph convolutional network (GCN) and gated recurrent unit (GRU) network, it learned spatio-spectral features. The fused latent features were able to provide more complementary and discriminative spatio-temporal-spectral fusion features for EEG signal representation, resulting in excellent recognition performance.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Xinyue Zhong, Yun Gu, Yutong Luo, Xiaomei Zeng, Guangyuan Liu
Summary: This study proposes a novel bi-hemispheric asymmetric attention network (Bi-AAN) for EEG-based emotion recognition. By combining a transformer structure with the asymmetric property of the brain's emotional response, the proposed method improves the recognition performance.
APPLIED INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Muzaffer Aslan
Summary: This study proposes a deep learning method based on GoogLeNet to automatically detect human emotions using EEG signals. By converting EEG signals into images and performing feature extraction using a pre-trained deep learning model, and applying the obtained features to machine learning algorithms for emotion classification, this method achieves high accuracy in emotion detection.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Physics, Multidisciplinary
Xin Zuo, Chi Zhang, Timo Hamalainen, Hanbing Gao, Yu Fu, Fengyu Cong
Summary: This study proposes a cross-subject emotion recognition framework based on EEG, using fused entropy features and a BiLSTM network. The research found that multi-scale entropy (MSE) is more effective than other single-entropy features in recognizing emotions. The performance of BiLSTM is further improved with the use of fused entropy features compared to single-type features.
Article
Computer Science, Information Systems
Musa Aslan, Muhammet Baykara, Talha Burak Alakus
Summary: Emotion recognition technology, utilizing electroencephalography signals, has been widely applied in various fields. This study conducted emotion analysis using EEG signals and identified the most impactful brain areas. Three different data sets were utilized, and the performances of these data sets were evaluated based on the brain areas. The study consisted of four stages: EEG data acquisition, preprocessing and feature extraction, deep learning model definition, and classification. The results showed high accuracy values for the data sets, with certain brain areas demonstrating more success and effectiveness.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Multidisciplinary Sciences
Badar Almarri, Sanguthevar Rajasekaran, Chun-Hsi Huang
Summary: This paper introduces a subject-independent emotion recognition framework that reduces subject-to-subject variability by adequate preprocessing, transforming, and feature extraction prior to analyzing emotion data. By utilizing unsupervised algorithms and support vector machine, the study outperforms other subject-independent studies in accurately classifying human affection based on EEG benchmarks.
Article
Computer Science, Theory & Methods
Yazhou Zhang, Dawei Song, Peng Zhang, Panpan Wang, Jingfei Li, Xiang Li, Benyou Wang
THEORETICAL COMPUTER SCIENCE
(2018)
Article
Computer Science, Artificial Intelligence
Yazhou Zhang, Dawei Song, Peng Zhang, Xiang Li, Panpan Wang
APPLIED INTELLIGENCE
(2019)
Article
Neurosciences
Xiang Li, Zhigang Zhao, Dawei Song, Yazhou Zhang, Jingshan Pan, Lu Wu, Jidong Huo, Chunyang Mu, Di Wang
FRONTIERS IN NEUROSCIENCE
(2020)
Article
Computer Science, Artificial Intelligence
Yazhou Zhang, Dawei Song, Xiang Li, Peng Zhang, Panpan Wang, Lu Rong, Guangliang Yu, Bo Wang
INFORMATION FUSION
(2020)
Article
Computer Science, Artificial Intelligence
Yazhou Zhang, Prayag Tiwari, Dawei Song, Xiaoliu Mao, Panpan Wang, Xiang Li, Hari Mohan Pandey
Summary: Conversational sentiment analysis is an emerging and challenging task that aims to discover the affective state and sentimental change of each person in a conversation. Current sentiment analysis approaches are inadequate for this subtask due to the lack of benchmark datasets and the difficulty in modeling interactions between individuals.
Article
Computer Science, Artificial Intelligence
Yazhou Zhang, Yaochen Liu, Qiuchi Li, Prayag Tiwari, Benyou Wang, Yuhua Li, Hari Mohan Pandey, Peng Zhang, Dawei Song
Summary: Sarcasm detection in conversation is a challenging artificial intelligence task that aims to discover ironically, contemptuously, and metaphorically implied information in daily conversations. A complex-valued fuzzy network leveraging quantum theory and fuzzy logic is proposed to address the intrinsic vagueness and uncertainty of human language in emotional expression and understanding, outperforming strong baselines in extensive experiments.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2021)
Article
Computer Science, Information Systems
Yazhou Zhang, Dan Ma, Prayag Tiwari, Chen Zhang, Mehedi Masud, Mohammad Shorfuzzaman, Dawei Song
Summary: Computational Linguistics (CL) associated with the Internet of Multimedia Things (IoMT) faces challenges in real-time speech understanding, deep fake video detection, emotion recognition, and home automation. The emergence of machine translation has greatly expanded CL solutions for natural language processing (NLP) applications. Sarcasm detection, a recent AI and NLP task, aims to uncover sarcastic information in texts generated in the IoMT. However, existing approaches often overlook the stance behind texts, which limits their effectiveness. In this research, we propose a new task called stance-level sarcasm detection (SLSD) and introduce an integral framework using BERT and a novel stance-centered graph attention network (SCGAT). Experimental results demonstrate the effectiveness of the SCGAT framework over state-of-the-art baselines.
ACM TRANSACTIONS ON INTERNET TECHNOLOGY
(2023)
Article
Computer Science, Information Systems
Yazhou Zhang, Prayag Tiwari, Lu Rong, Rui Chen, Nojoom A. Alnajem, M. Shamim Hossain
Summary: The recent development of artificial intelligence applications has led to the creation of a large number of multi-modal records of human communication, which contain latent subjective attitudes and opinions. Sentiment and emotion analysis is of great value in improving affective services. Finding an optimal way to learn people's sentiments and emotional representations has been a challenging problem. In this work, a multi-task representation learning network called KAMT is proposed to address the challenges of multi-modal fused representation and the interaction between sentiment and emotion.
ACM TRANSACTIONS ON MULTIMEDIA COMPUTING COMMUNICATIONS AND APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Yazhou Zhang, Jinglin Wang, Yaochen Liu, Lu Rong, Qian Zheng, Dawei Song, Prayag Tiwari, Jing Qin
Summary: Sarcasm, sentiment, and emotion are closely related in the research of artificial intelligence and affective computing. This paper proposes a multimodal multitask learning model (M2Seq2Seq) to address the challenges of context dependency, multimodal fusion, and multitask interaction. The model incorporates encoder-decoder architecture and attention mechanisms to capture contextual dependency and multimodal interactions. Experimental results demonstrate the effectiveness of M2Seq2Seq over state-of-the-art baselines.
INFORMATION FUSION
(2023)
Article
Engineering, Civil
Yazhou Zhang, Prayag Tiwari, Qian Zheng, Abdulmotaleb El Saddik, M. Shamim Hossain
Summary: Traffic events are a major cause of traffic accidents, and detecting these events poses a challenge in traffic management and intelligent transportation systems (ITSs). This paper proposes a multimodal coupled graph attention network (MCGAT) that extracts valuable information from various traffic data sources and represents it in a graphical structure. The proposed model outperforms state-of-the-art baselines in terms of F1 and accuracy, with significant improvements.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Proceedings Paper
Computer Science, Theory & Methods
Panpan Wang, Zhao Li, Yazhou Zhang, Yuexian Hou, Liangzhu Ge
PROCEEDINGS OF THE 28TH ACM INTERNATIONAL CONFERENCE ON INFORMATION & KNOWLEDGE MANAGEMENT (CIKM '19)
(2019)
Proceedings Paper
Computer Science, Information Systems
Yazhou Zhang, Dawei Song, Xiang Li, Peng Zhang
ADVANCES IN INFORMATION RETRIEVAL (ECIR 2018)
(2018)