Affect Detection for Human-Horse Interaction

Abstract

In this work, we aim to study the potential use of affect recognition techniques for examining the interaction between humans and horses using qualitative and quantitative methods. To this end, we propose a multi-modal portable system for physiological signal acquisition such as the electrocardiogram (ECG), electromyogram (EMG), and electroencephalogram (EEG). The proposed system is used to acquire signals while users are interacting with horses. The captured signals will then be used in order to quantitatively evaluate human and equine interaction by mapping the signals to the emotional state of the subjects using machine learning techniques. In this preliminary study, ECG based features were utilised in order to create a supervised classification model that can identify emotions elicited during human-horse interaction. Experimental results provide evidence about the efficiency of the proposed approach in distinguishing between negative and positive emotions, reaching a classification accuracy of 74.21%.

Examining Human-Horse Interaction by Means of Affect Recognition via Physiological Signals

“…In their previous work [36], the authors of this paper examined the use of ECG-based features in order to distinguish between negative and positive emotion during human-horse interaction. The experimental evaluation provided promising results on the efficiency of ECG signals in distinguishing between negative and positive emotions, reaching a classification accuracy of 74.21% [36]. However, it is clear that further research on human-horse interaction using various physiological signals is needed in order to validate the feasibility of using such signals in the field of human-horse interaction.…”

Driver Stress Detection from Physiological Signals by Virtual Reality Simulator

“…Emotional sensing systems offer the potential to identify them and be able to address the causal factors and are useful in applications in multiple domains, especially those focused on health, such as stress [16,38,42] or other types of emotions [3,5,31,43,44], where physical and mental states can be monitored in real time [2,24,36] and can act accordingly [28] as intelligent assistants [21,45]; in environmental assisted living [27,33]; in the industries of games [26], robots [46,47], domotics [48], marketing, or recommendations [4,34,37,49]; and in the study of social behavior [23,30,34], authentication and security [18,39,48], or education [49], among others.…”

Machine learning-based affect detection within the context of human–horse interaction

“…• ECG: 84 ECG-based features, commonly used in affect recognition studies [2,3,35,36], were extracted using the Augsburg Biosignal Toolbox (AuBT) [43]. [35,36], were extracted using the Augsburg Biosignal Toolbox (AuBT) [43].…”

“…• ECG: 84 ECG-based features, commonly used in affect recognition studies [2,3,35,36], were extracted using the Augsburg Biosignal Toolbox (AuBT) [43]. [35,36], were extracted using the Augsburg Biosignal Toolbox (AuBT) [43]. The following 7 features were extracted from each of the raw EMG signal, its first derivative and its second derivative: mean, median, standard deviation, minima, maxima, and the number of times per time unit that the signal reached the minima and the maxima.…”

“…In another study, Lanata et al [33] examined the use of ECG signals for assessing human-horse interaction before the interaction, during visual-olfactory interaction and while grooming the horse, with the purpose of using machine learning to detect the interaction activity, achieving an accuracy of 70.87% using the Nearest Mean classifier and 90.95% using an SVM classifier on the same data during a later study [34]. Machine learning approaches and ECG-based features were also employed by Althobaiti et al [35] for distinguishing between negative and positive emotions during human-horse interaction, reporting an accuracy of 74.21%. In another study, Althobaiti et al [36] examined the use of ECG, EMG, and EEG-based features in combination with various classification algorithms for the detection of the Valence and Arousal dimensions of the emotional state of humans during human-horse interaction.…”