Pawlytics: Pet Behaviour and Health Analysis Using AI / ML

This paper presented Pawlytics, an AI-based framework for pet emotion recognition and behavioural analysis aimed at supporting proactive pet healthcare management. The proposed framework employs a lightweight MobileNetV2 deep learning model to classify pet emotions and generate behavioural insights and personalized recommendations for pet owners and veterinary professionals. Experimental evaluation demonstrated that the proposed framework achieved an overall classification accuracy of 71%, with competitive Precision, Recall, and F1-Score values across all emotional classes. The results indicate that lightweight deep learning architectures can effectively perform multi-class pet emotion classification while maintaining computational efficiency suitable for real-time mobile deployment. Comparative analysis with existing approaches showed that Pawlytics provides a more comprehensive solution by integrating emotion recognition, behavioural analytics, recommendation mechanisms, and veterinary decision support within a single framework. Future work will focus on incorporating multimodal emotion recognition using audio and video information, integrating wearable sensors for continuous monitoring, developing disease prediction models, and exploring transformer-based architectures and larger datasets to improve classification performance and generalization capability. These enhancements have the potential to evolve Pawlytics into a comprehensive next-generation intelligent pet healthcare platform.

[1]          V. Tanwar, "Understanding Dog Emotions Through Deep Learning: A CNN-Based Classification Framework," IEEE Access, vol. 12, pp. 45231-45240, 2024. [2]          M. Lee, J. Hernandez, and K. Park, "AI-Based Behaviour Monitoring in Pets Using Machine Learning Techniques," IEEE Transactions on Artificial Intelligence, vol. 4, no. 3, pp. 221-230, 2023. [3]          A. Patel, R. Sharma, and P. Gupta, "Machine Learning in Veterinary Healthcare: A Review of Predictive Approaches," Springer Journal of Veterinary Informatics, vol. 18, no. 2, pp. 101-115, 2022. [4]          K. Ayyaswamy, S. Kumar, and M. Prakash, "AI-Based Healthcare Systems for Pets and Birds: Opportunities and Challenges," Biomedical Research International, vol. 2024, Art. no. 9876543, pp. 1-12, 2024. [5]          Y. Chen, L. Wang, and J. Liu, "IoT-Enabled Pet Health Tracking Using Wearable Sensors and Machine Learning," Internet of Things, vol. 15, Art. no. 100415, 2021. [6]          R. Singh, P. Verma, and A. Kaur, "Deep Learning for Animal Activity Recognition Using Convolutional Neural Networks," in Proc. IEEE Int. Conf. Computational Intelligence, 2022, pp. 112-118. [7]          K. Johnson and M. Carter, "Predictive Analytics in Pet Care: Machine Learning Approaches for Preventive Healthcare," Int. J. Veterinary Science and Technology, vol. 10, no. 4, pp. 145-153, 2021. [8]          X. Li, Y. Zhao, and H. Chen, "Computer Vision Techniques for Animal Behaviour Analysis," Pattern Recognition Letters, vol. 136, pp. 152-160, 2020. [9]          L. Wang and S. Huang, "Cloud-Based Pet Health Records for Smart Veterinary Services," Journal of Cloud Computing, vol. 10, no. 1, pp. 1-11, 2021. [10]       H. Guo, Z. Liu, and Y. Zhang, "Multimodal Pet Emotion Recognition Using Deep Learning," IEEE Access, vol. 11, pp. 65421-65433, 2023. [11]       M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L. Chen, "MobileNetV2: Inverted Residuals and Linear Bottlenecks," in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), Salt Lake City, UT, USA, Jun. 2018, pp. 4510-4520.