TY - CONF A1 - Kilic, Özge T1 - Tracking collective behavioural responses of termites with machine learning pose estimation software N2 - Termites, as eusocial insects, exhibit complex social behaviors critical to their ecological roles. This study employs cutting-edge machine learning techniques, specifically SLEAP pose estimation software, to investigate the interactions of *Reticulitermes flavipes* within controlled environments. By leveraging advanced tracking methods, this research addresses challenges such as track breaks in movement data and uses custom coding solutions to ensure accurate reconnections, preserving the continuity and reliability of individual and group behavior analyses. Custom-designed 3D-printed habitats facilitate controlled observations, while machine learning enables the annotation and tracking of key body segments (mandible, thorax, and abdomen). This approach provides granular insights into colony dynamics, highlighting task differentiation and social clustering. Future applications include the development of automated, continuous recording setups and expanded experiments to examine the impact of varying environmental conditions. This interdisciplinary work, combining biology and data science, advances our understanding of termite social systems and paves the way for predictive simulations of their activity. Such insights contribute to ecological management strategies and deepen our knowledge of collective animal behaviors. T2 - ASAB 2024 CY - Exeter, United Kingdom DA - 21.04.2024 KW - Machine learning KW - Artifical Intelligent KW - Social behaviour KW - 3D Printing KW - Pose estimation KW - Tracking PY - 2024 AN - OPUS4-61830 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -