Automated recognition of emotional states of horses from facial expressions.
- Journal Article
Summary
The research article discusses a groundbreaking study that uses artificial intelligence (AI) models to recognize and interpret the emotional states of horses through their facial expressions. This is the first of its kind in a growing field known as animal affective computing.
Overview of the Research
The researchers wanted to understand and interpret the emotional states of horses. Prior studies in the field of animal affective computing primarily focused on identifying pain in animals. This study, however, aimed to delineate from that approach and explore emotions beyond pain in horses. Their primary tool was AI models capable of interpreting horses’ facial expressions.
Research Method and Techniques
Two different AI approaches were implemented to recognize the horses’ emotional states:
- Deep Learning Model: This model makes use of video footage as its data source.
- Machine Learning Model: This model utilizes EquiFACS (Equine Facial Action Coding System) annotations for data.
Comparison between Two Pipeline Approaches
The two models were compared and evaluated based on their performance. The Deep Learning model, which used video footage, had better performance compared to the Machine Learning model that used EquiFACS annotations.
Emotional States Detectable
The AI models aimed to distinguish four different emotional states:
- Baseline
- Positive Anticipation
- Disappointment
- Frustration
Results of the Study
The deep learning AI model could classify the four emotional states with an accuracy of 76%. However, it found it more challenging to distinguish between ‘anticipation’ and ‘frustration’, and only achieved 61% success in correctly distinguishing these emotions.
Cite This Article
Publication
Researcher Affiliations
- Information Systems Department, University of Haifa, Haifa, Israel.
- Computer Science Department, University of Haifa, Haifa, Israel.
- Information Systems Department, University of Haifa, Haifa, Israel.
- Information Systems Department, University of Haifa, Haifa, Israel.
- Information Systems Department, University of Haifa, Haifa, Israel.
- Faculty of Electrical Engineering, Technion, Israel Institute of Technology, Haifa, Israel.
- Faculty of Electrical Engineering, Technion, Israel Institute of Technology, Haifa, Israel.
- Information Systems Department, University of Haifa, Haifa, Israel.
- Department of Life Sciences, Joseph Banks Laboratories, University of Lincoln, Lincoln, United Kingdom.
- Information Systems Department, University of Haifa, Haifa, Israel.
MeSH Terms
- Horses / psychology
- Animals
- Facial Expression
- Emotions / physiology
- Machine Learning
- Deep Learning
- Male
- Humans
Conflict of Interest Statement
Citations
This article has been cited 5 times.- Guo X, Shi L, Ma B, Feng C, Liu Z. Research on improved models for facial expression recognition in mice with abnormal glucose metabolism. Sci Rep 2026 Feb 10;16(1).
- Bhave A, Kieson E, Hafner A, Gloor PA. Identifying Novel Emotions and Wellbeing of Horses from Videos Through Unsupervised Learning. Sensors (Basel) 2025 Jan 31;25(3).
- O'Connell E, Dyson S, McLean A, McGreevy P. No More Evasion: Redefining Conflict Behaviour in Human-Horse Interactions. Animals (Basel) 2025 Jan 31;15(3).
- Feighelstein M, Ricci-Bonot C, Hasan H, Weinberg H, Rettig T, Segal M, Distelfeld T, Shimshoni I, Mills DS, Zamansky A. Correction: Automated recognition of emotional states of horses from facial expressions. PLoS One 2025;20(2):e0319501.
- König von Borstel U, Kienapfel K, McLean A, Wilkins C, McGreevy P. Hyperflexing the horse's neck: a systematic review and meta-analysis. Sci Rep 2024 Oct 2;14(1):22886.