A Survey on Physiological Emotion Recognition

Abstract

Every human has disparate emotions in their daily life, and these emotions are not fully apprehended. In this paper, we are dis- cussing various types of emotion recognition methods. We have two kinds of signals, such as physiological and non-physiological signals. We mostly use physiological signals like electrocardiograms, galvanic skin respiration, temperature, electroencephalograms, electromyograms, and so on. One of the main drawbacks of these papers is that they cannot predict human absolute emotions.

Introduction

Emotions are a fundamental part of human life, helping individuals regulate their minds and bodies, interact socially, make decisions, and navigate daily experiences. Emotion is generally explained through three components: physiological, neurological, and cognitive processes, where bodily reactions and brain activity together shape emotional responses.

The text reviews several research works on emotion recognition, especially using physiological and behavioral signals with machine learning and deep learning techniques.

Key emotion models discussed include:

• Ekman’s Model, which identifies six universal emotions: happiness, sadness, surprise, fear, anger, and disgust.

• Plutchik’s Model, which presents eight basic emotions arranged in a wheel, highlighting emotional intensity and opposites.

• Russell’s Model, which represents emotions along two dimensions: valence (pleasant–unpleasant) and arousal (high–low activation).

Multiple datasets such as DEAP, MAHNOB-HCI, AMIGOS, DREAMER, and SEED are widely used for emotion recognition experiments involving EEG, physiological, audio, and video data.

The reviewed studies apply techniques like CNNs, DCNNs, LSTM, SVM, Naive Bayes, and feature selection methods to classify emotions, often focusing on arousal and valence levels. Physiological signals (EEG, ECG, heart rate, skin temperature, electrodermal activity) are emphasized as reliable indicators of emotion, especially when facial expressions or speech are unreliable.

Recent research also explores multimodal and mobile-based emotion recognition, combining behavioral and physiological signals through wearable devices, though limitations include small datasets and a narrow range of emotions.

From the comparative analysis, Support Vector Machines (SVM) emerge as one of the most commonly used and effective algorithms across studies. Overall, accurate emotion recognition depends heavily on well-defined features, robust datasets, and suitable algorithm selection, with fused physiological signals showing strong potential for improved performance.

If you want, I can also:

• Shorten this further for exam notes

• Rewrite it in simpler language

• Turn it into an abstract or conclusion section

Conclusion

Emotions are classified into two types as primary and secondary emotions. Primary emotions which consist of Sad, Happy, Fear, Disgust etc..., and in Secondary emotions which consist of Envy, Pride, Jealous etc... Here we go through different issues such as gestures, speech ,mental state all are varying. A well founded detector is not equipped for correct emotion detection .Here by using Hybrid particle swarm optimization and Support vector machine algorithm checking for its efficient tool.

References

[1] Tongshuai Song, Guanming Lu*, Jingjie Yan, \"Emotion Recognition Based on Physiological Signals Using Convolution Neural Networks \", ICMLC 2020, February 15–17, 2020. [2] Luz SantaMaria-Granados , Mario Munoz- Organero , Gustavo Ramirez-Gonzalez , Enas Abdulhay , And N. Arun Kumar, Using Deep Con- volutional Neural Network for Emotion Detection on a Physiological Signals Dataset (AMIGOS) \"IEEE Transactions and Journals, Vol. 4, 2016. [3] Salma Alhagry , Aly Aly Fahmy, Reda A. El- Khoribi,\"Emotion Recognition based on EEG us- ing LSTM Recurrent Neural Network\", International Journal of Advanced Computer Science and Applications, Vol. 8, No.10,2017 [4] Kangning Yang, Chaofan Wang, Yue Gu, Zhanna Sarsenbayeva, Benjamin Tag, Tilman Dingler, Greg Wadley, and Jorge Goncalves, \"Behavioral and Physiological Signals-Based Deep Multimodal Approach for Mobile Emotion Recognition\",IEEE Transactions on Affective Computing,2015. [5] Diego Fabiano and Shaun Canavan, “Emotion Recognition Using Fused Physiological Signals”. IEEE, 78-1-7281-3888-6/19/2019

Copyright

Copyright © 2026 Prof. Viji C. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.