Automatic Recognition of Multiple Emotional Classes from EEG Signals through the Use of Graph Theory and Convolutional Neural Networks

Sensors, Год журнала: 2024, Номер 24(18), С. 5883 - 5883

Опубликована: Сен. 10, 2024

Emotion is a complex state caused by the functioning of human brain in relation to various events, for which there no scientific definition. recognition traditionally conducted psychologists and experts based on facial expressions—the traditional way recognize something limited associated with errors. This study presents new automatic method using electroencephalogram (EEG) signals combining graph theory convolutional networks emotion recognition. In proposed model, firstly, comprehensive database musical stimuli provided induce two three emotional classes, including positive, negative, neutral emotions. Generative adversarial (GANs) are used supplement recorded data, then input into suggested deep network feature extraction classification. The can extract dynamic information from EEG data an optimal manner has 4 GConv layers. accuracy categorization classes respectively, 99% 98%, according strategy. model been compared recent research algorithms promising results. be complete brain-computer-interface (BCI) systems puzzle.

Язык: Английский

---

Introducing a Novel Dataset for Facial Emotion Recognition and Demonstrating Significant Enhancements in Deep Learning Performance through Pre-processing Techniques

Heliyon, Год журнала: 2024, Номер 10(20), С. e38913 - e38913

Опубликована: Окт. 1, 2024

Язык: Английский

---

An Effective Methodology for Diabetes Prediction in the Case of Class Imbalance

Bioengineering, Год журнала: 2025, Номер 12(1), С. 35 - 35

Опубликована: Янв. 6, 2025

Diabetes causes an increase in the level of blood sugar, which leads to damage various parts human body. data are used not only for providing a deeper understanding treatment mechanisms but also predicting probability that one might become sick. This paper proposes novel methodology perform classification case heavy class imbalance, as observed PIMA diabetes dataset. The proposed uses two steps, namely resampling and random shuffling prior defining model. is tested with versions cross validation appropriate cases imbalance-k-fold stratified k-fold validation. Our findings suggest when having imbalanced data, randomly train/test split can help improve estimation metrics. outperform existing machine learning algorithms complex deep models. Applying our simple fast way predict labels imbalance. It does require additional techniques balance classes. involve preselecting important variables, saves time makes model easy analysis. it effective initial further modeling Moreover, methodologies show how effectiveness models based on standard approaches make them more reliable.

Язык: Английский

---

CNN-LSTM based emotion recognition using Chebyshev moment and K-fold validation with multi-library SVM

PLoS ONE, Год журнала: 2025, Номер 20(4), С. e0320058 - e0320058

Опубликована: Апрель 7, 2025

Human emotions are not necessarily tends to produce right facial expressions as there is no well defined connection between them. Although, human spontaneous, their depend a lot on mental and psychological capacity either hide it or show explicitly. Over decade, Machine Learning Neural Networks methodologies most widely used by the researchers tackle these challenges, deliver an improved performance with accuracy. This paper focuses analyzing driver’s determine mood emotional state while driving ensure safety. we propose hybrid CNN-LSTM model in which RESNET152 CNN along Multi-Library Support Vector for classification purposes. For betterment of feature extraction, this study has considered Chebyshev moment plays important role repetition process gain primary features K-fold validation helps evaluate models terms both training, loss, was evaluated compared existing approaches like CNN-SVM ANN-LSTM where proposed delivered better results than other considered.

Язык: Английский

---

Integrating end-to-end multimodal deep learning and domain adaptation for robust facial expression recognition

Image and Vision Computing, Год журнала: 2025, Номер 159, С. 105548 - 105548

Опубликована: Апрель 30, 2025

Язык: Английский

---

Evaluating the capacity of large language models to interpret emotions in images

PLoS ONE, Год журнала: 2025, Номер 20(6), С. e0324127 - e0324127

Опубликована: Июнь 3, 2025

The integration of artificial intelligence, specifically large language models (LLMs), in emotional stimulus selection and validation offers a promising avenue for enhancing emotion comprehension frameworks. Traditional methods this domain are often labor-intensive susceptible to biases, highlighting the need more efficient scalable alternatives. This study evaluates capability GPT-4, recognizing rating emotions from visual stimuli, focusing on two primary dimensions: valence (positive, neutral, or negative) arousal (calm, stimulated). By comparing performance GPT-4 against human evaluations using well-established Geneva Affective PicturE Database (GAPED), we aim assess model's efficacy as tool automating elicitation stimuli. Our findings indicate that closely approximates ratings under zero-shot learning conditions, although it encounters some difficulties accurately classifying subtler cues. These results underscore potential LLMs streamline process, thereby reducing time labor associated with traditional methods.

Язык: Английский

---

Emotion Recognition Using EEG Signals through the Design of a Dry Electrode Based on the Combination of Type 2 Fuzzy Sets and Deep Convolutional Graph Networks

Biomimetics, Год журнала: 2024, Номер 9(9), С. 562 - 562

Опубликована: Сен. 18, 2024

Emotion is an intricate cognitive state that, when identified, can serve as a crucial component of the brain-computer interface. This study examines identification two categories positive and negative emotions through development implementation dry electrode electroencephalogram (EEG). To achieve this objective, EEG created using silver-copper sintering technique, which assessed Scanning Electron Microscope (SEM) Energy Dispersive X-ray Analysis (EDXA) evaluations. Subsequently, database generated utilizing designated electrode, based on musical stimulus. The collected data are fed into improved deep network for automatic feature selection/extraction classification. architecture structured by combining type 2 fuzzy sets (FT2) convolutional graph networks. fabricated demonstrated superior performance, efficiency, affordability compared to other electrodes (both wet dry) in study. Furthermore, was examined noisy environments robust resistance across diverse range Signal-To-Noise ratios (SNRs). proposed model achieved classification accuracy 99% distinguishing between emotions, improvement approximately 2% over previous studies. manufactured very economical cost-effective terms manufacturing costs recent network, combined with be used real-time applications long-term recordings that do not require gel.

Язык: Английский

---

Asymmetric measures of polar Chebyshev chaotic map for discrete/dimensional emotion recognition using PPG

Biomedical Signal Processing and Control, Год журнала: 2024, Номер 100, С. 107089 - 107089

Опубликована: Окт. 21, 2024

Язык: Английский

---

Swin Transformer with Enhanced Dropout and Layer-wise Unfreezing for Facial Expression Recognition in Mental Health Detection

Engineering Technology & Applied Science Research, Год журнала: 2024, Номер 14(6), С. 19016 - 19023

Опубликована: Дек. 2, 2024

This study presents an improved Facial Expression Recognition (FER) model using Swin transformers for enhanced performance in detecting mental health through facial emotion analysis. In addition, some techniques involving better dropout and layer-wise unfreezing were implemented to reduce overfitting. evaluates the proposed models on benchmark datasets such as FER2013 CK+ real-time Genius HR data. Model A has no layer, B focal loss, C unfreezing. was best among all models, achieving test accuracies of 71.23% 78.65% CK+. Weighted cross-entropy loss image augmentation used handle class imbalance. Based predictions, a scoring mechanism designed analyze employees' next 30 days. The higher score, risk health. demonstrates practical version transformer FER early intervention.

Язык: Английский

---

Efficient Facial Expression Recognition Through Lightweight CNN Technique on Public Datasets

SN Computer Science, Год журнала: 2024, Номер 6(1)

Опубликована: Дек. 19, 2024

Язык: Английский

---