| Title |
Anomaly detection approach for cow behavior without relying on a predefined gold standard |
| Authors |
Nakrošis, Arnas ; Kaedbey, Jad ; Klepachevskyi, Dmytro ; Klumbys, Marius ; Aliyev, Vusal ; Paulauskaitė-Tarasevičienė, Agnė |
| Full Text |
|
| Is Part of |
CEUR workshop proceedings: IVUS 2026: proceedings of the 31st international conference on information society and university studies (IVUS 2026), Kaunas, Lithuania, May 12, 2026 / edited by Audrius Lopata,Tomas Krilavičius, Ilona Veitaitė.. Aachen : CEUR-WS. 2026, vol. 4317, p. 1-11.. ISSN 1613-0073 |
| Keywords [eng] |
Livestock Monitoring ; Thermal Imaging ; Deep Learning ; Unsupervised Learning ; Anomaly Detection |
| Abstract [eng] |
thermal video data. Detecting abnormal behavior in livestock is particularly challenging due to the absence of labeled anomalies, high intra-animal variability, and environmental noise in thermal recordings. Four unsupervised learning models – K-means clustering, Isolation Forest, Convolutional Autoencoder, and Generative Adversarial Network, were implemented, evaluated, and comparatively analyzed with respect to their effectiveness in detecting abnormal behavioral patterns. The results show that unsupervised methods combined with thermal imaging can provide an effective solution for automated livestock monitoring. Despite the inherent complexity of the task, the Convolutional Autoencoder achieved the highest performance, attaining an accuracy of 86.53%, thereby underscoring the potential of deep unsupervised models and the need for further investigation. |
| Published |
Aachen : CEUR-WS |
| Type |
Conference paper |
| Language |
English |
| Publication date |
2026 |
| CC license |
|
