Behavior Analysis Archives - Tech4Animals

The Art of Chimpanzee (Pan troglodytes) Diplomacy: Unraveling the secrets of successful negotiations using machine learning

annazam — Mon, 21 Jul 2025 11:47:18 +0000

This paper introduces a novel segment-based framework for enhancing the explainability of deep learning models in animal affective computing, a field that uses technology to understand animal emotions. The core challenge addressed is the “black box” nature of deep learning, which makes it difficult to understand why a model makes a particular decision, hindering trust and adoption among researchers. The proposed framework focuses on evaluating visual explanations, specifically saliency maps, by quantifiably assessing how well they align with biologically meaningful semantic parts of an animal, such as eyes, ears, and mouth. Through case studies on cat pain, horse pain, and dog emotions, the authors demonstrate how this framework can provide quantifiable insights into which facial regions are most important for a classifier’s decision, revealing that the eye area consistently holds the most significance. Ultimately, this framework aims to bridge the gap between complex AI models and expert knowledge, fostering both validation of known indicators and the discovery of new ones in animal behavior.

The post The Art of Chimpanzee (Pan troglodytes) Diplomacy: Unraveling the secrets of successful negotiations using machine learning appeared first on Tech4Animals.

Automated Landmark-based Cat Facial Analysis and its Applications

annazam — Sat, 03 Aug 2024 06:49:35 +0000

This paper systematically explores the utility of an automated 48-landmark detector for cat facial analysis, addressing the labor-intensive nature of manual annotation. The study developed AI pipelines for three benchmark tasks using two previously collected datasets: cat breed recognition, cephalic type recognition, and pain recognition. While replacing manual landmarks with automated ones generally decreased performance across all tasks, this reduction was an acceptable trade-off for full automation in some areas. Specifically, automated pipelines achieved 75% accuracy in cephalic type recognition and 66% in pain recognition, suggesting landmark-based approaches show promise for automated pain assessment and morphological explorations. However, the breed recognition pipeline performed poorly, indicating that deep learning approaches using direct image data might be more suitable for this task. The study emphasizes the need for valid benchmarks and publicly accessible datasets in animal facial analysis, similar to human affective computing, while also considering ethical implications.

The post Automated Landmark-based Cat Facial Analysis and its Applications appeared first on Tech4Animals.

Digitally-Enhanced Dog Behavioral Testing: Getting Help from the Machine

annazam — Wed, 16 Aug 2023 20:50:35 +0000

This study introduces a novel paradigm for digitally-enhanced dog behavioral testing, aiming to provide a more objective, robust, and resource-efficient alternative to traditional subjective and time-consuming methods like questionnaires or observation. Researchers tested 53 dogs using a “Stranger Test” protocol, collecting data from three expert behavioral scores and owner-completed Canine Behavioral Assessment and Research Questionnaires (C-BARQ). Utilizing an automated computational approach, specifically the BLYZER system, dog trajectories were extracted and clustered in a “human-free” way from video footage. The unsupervised clustering revealed two distinct behavioral profiles, showing a significant difference in stranger-directed fear C-BARQ scores and effectively separating relaxed dogs from those exhibiting excessive behaviors based on expert evaluations. Furthermore, a machine learning classifier achieved an accuracy of 78% in predicting expert scores of dog coping styles, and a regression model predicted C-BARQ categories with varying success, performing best for Owner-Directed Aggression and Excitability. This research highlights the significant value and promise of AI in the field of dog behavioral assessment, offering a path towards more reliable and objective evaluations.

The post Digitally-Enhanced Dog Behavioral Testing: Getting Help from the Machine appeared first on Tech4Animals.

Clustering for Automated Exploratory Pattern Discovery in Animal Behavioral Data

annazam — Wed, 17 Aug 2022 08:05:29 +0000

This paper introduces a method to automatically identify hidden behavioral patterns in animals using unsupervised machine learning. It leverages k-means clustering and PCA on automatically extracted movement features from dog video trials to overcome human bias and labor-intensive traditional methods. The study successfully applied this to a “Stranger Test,” showing that clusters derived from objective movement features, such as intensity of use and contact ratio, correlated with owner-reported stranger fear and aggression (C-BARQ scores), thus enabling exploratory insights prior to hypothesis formation. However, the results are preliminary due to a small sample size and require further validation.

The post Clustering for Automated Exploratory Pattern Discovery in Animal Behavioral Data appeared first on Tech4Animals.

Objective Video-based Assessment of ADHD-like Canine Behavior using Machine Learning

annazam — Fri, 20 Aug 2021 12:32:17 +0000

This paper introduces a novel objective video-based method for assessing canine Attention-Deficit/Hyperactivity Disorder (ADHD)-like behavior using machine learning. The study trained a Random Forest classifier on video recordings of dogs in a veterinary consultation room, analyzing movement features such as total distance and average speed. This method achieved 81% accuracy in distinguishing between dogs with clinically treated ADHD-like behavior and healthy controls. The resulting H-score, which quantifies the degree of ADHD-like behavior, showed a reduction in 8 out of 11 patients following medical treatment. Behavioral experts perceived the H-score as a valuable, objective tool to support clinical assessment and communication with owners, although they emphasized it should not be the sole basis for diagnosis.

The post Objective Video-based Assessment of ADHD-like Canine Behavior using Machine Learning appeared first on Tech4Animals.

Towards a Methodology for Data-Driven Automatic Analysis of Animal Behavioral Patterns

annazam — Thu, 20 Aug 2020 12:44:59 +0000

This paper addresses the significant challenge in animal-related disciplines of effectively measuring behavior, particularly with the advent of computational approaches that generate vast amounts of data and offer practically endless measurement possibilities. The authors, including Tom Menaker and Anna Zamansky, propose a data-driven framework called Data-Driven Behavioral Pattern Analysis (DD-BPA). This framework aims to guide researchers in selecting relevant behavioral parameters by applying data mining techniques to extract insights from experimental data, shifting the focus from merely automating manual coding to actively supporting data analysis decisions. They demonstrate their approach through a clustering-based analysis of animal trajectories from a dog facial preference experiment at the Clever Dog Lab, which effectively identified “prevalent areas of stay” for the animal subjects using tools like Blyzer.

The post Towards a Methodology for Data-Driven Automatic Analysis of Animal Behavioral Patterns appeared first on Tech4Animals.

Computational Analysis of Movement Patterns of Dogs with ADHD-like Behavior

annazam — Tue, 20 Aug 2019 12:53:04 +0000

This paper presents a study that uses a self-developed computational tool, K9-Blyzer, to objectively analyze movement patterns of dogs with ADHD-like behavior from video footage collected during veterinary consultations. The study compared 12 dogs diagnosed with ADHD-like behavior (H-group) to a control group of 12 dogs with no behavioral problems (C-group). The computational analysis identified seven significant differences in movement parameters between the groups, leading to the identification of three characteristic dimensions for dogs with ADHD-like behavior: high speed of movement, large coverage of room space, and frequent re-orientation in room space. These findings suggest that computational tools can potentially provide objective support for the assessment, diagnosis, and clinical treatment of ADHD-like behavior in dogs, complementing owner reports.

The post Computational Analysis of Movement Patterns of Dogs with ADHD-like Behavior appeared first on Tech4Animals.

What Does That Head Tilt Mean? Brain Lateralization and Sex Differences in the Processing of Familiar Human Speech by Domestic Dogs

annazam — Fri, 31 Oct 2025 08:23:30 +0000

This study explores the underlying causes of head-tilting behavior in domestic dogs during human-animal interactions. Researchers utilized manual coding and AI analysis to observe how 103 dogs responded to varying levels of communicative cues from their owners. The findings indicate that dogs tilt their heads significantly more when hearing familiar words and phrases compared to neutral or unfamiliar speech. Notably, the behavior showed a right-sided lateralization, suggesting that dogs process meaningful human language primarily in the left hemisphere of the brain. The data also revealed sex-based differences, as neutered males tilted their heads more frequently than spayed females, reflecting similar auditory processing patterns found in humans. Ultimately, the research suggests that the head tilt serves as a physical indicator of cognitive processing and lateralized brain activity in response to socially relevant speech.

The post What Does That Head Tilt Mean? Brain Lateralization and Sex Differences in the Processing of Familiar Human Speech by Domestic Dogs appeared first on Tech4Animals.

Close Encounters of the Cat Kind: The influence of context and sex on facial signaling proximity in domesticated cats (Felis silvestris catus)

annazam — Wed, 01 Apr 2026 12:42:59 +0000

Accurately measuring animal social bonds often relies on analyzing spatial proximity, yet advanced automated tracking methods have rarely been applied to feline interactions. To address this gap, this study utilized AI-based computer vision systems to extract precise distance data from video recordings of domestic cats engaging in social facial signaling. Contrary to initial predictions, the analysis revealed that cats maintained significantly closer physical proximity during non-affiliative encounters than during friendly, affiliative interactions. Additionally, the sex of the interacting pairs strongly influenced their spacing, with female-female dyads staying the closest across all contexts. Ultimately, this work demonstrates how machine learning tools can uncover nuanced behavioral strategies that animals use to navigate social environments.

The post Close Encounters of the Cat Kind: The influence of context and sex on facial signaling proximity in domesticated cats (Felis silvestris catus) appeared first on Tech4Animals.

Dog Facial Landmarks Detection and its Applications for Facial Analysis

annazam — Tue, 01 Jul 2025 12:47:06 +0000

Understanding canine emotions often relies on manual behavioral coding, a time-intensive process that is difficult to scale. To address this bottleneck, this study introduces a computer vision approach using 46 anatomy-based facial landmarks to automate the analysis of dog expressions. Researchers developed a dataset of over 3,700 annotated images and trained machine learning models to track subtle facial changes in video data. The findings demonstrate that this automated system successfully classifies distinct emotional states, such as positive anticipation and frustration, while accurately identifying rapid facial movements. Furthermore, utilizing this technology as a computer-assisted coding aid yielded a 41% reduction in the time required for human experts to manually annotate behavior.

The post Dog Facial Landmarks Detection and its Applications for Facial Analysis appeared first on Tech4Animals.