OA2075. Teaching safe falling: An AI-based educational approach to injury prevention

Abstract Background Falls remain a leading cause of injury, especially among vulnerable populations. Early educational interventions focusing on safe fall strategies can mitigate fall-related risks and injuries. The SAFE FALL project merges training, biomechanics, and artificial intelligence to create a more effective educational programme. By analysing students’ postural dynamics before and after intervention, the project seeks to objectively measure improvements in fall responses through advanced technological tools. Objectives This study aims to assess the effectiveness of fall-prevention educational interventions by capturing and analysing students’ body movements using cutting-edge AI-based biomechanical analysis methods. Specifically, it focuses on quantifying changes in fall strategies and postural adjustments induced by targeted training programmes. Results Students’ movements were recorded during fall simulations using high-resolution camera systems. Pose estimation algorithms based on deep learning techniques were employed to extract key anatomical landmarks. The segmentation of participants from the environment was achieved using the Segment Anything Model 2 (SAM2), allowing precise isolation of body dynamics. Furthermore, active motion tracking was implemented to monitor key biomechanical parameters such as centre of mass displacement, angular momentum, and joint kinematics during falls. Machine learning models were developed to analyse pre- and post-intervention differences, revealing marked improvements in protective movement strategies, balance recovery, and fall mitigation behaviours. Conclusions The integration of AI-driven pose estimation, segmentation, and motion tracking tools offers a robust framework for evaluating fall prevention training. The SAFE FALL project demonstrates that combining education with advanced technological assessment provides measurable benefits, facilitating better-designed interventions and potentially reducing fall risks. Key messages • Leveraging artificial intelligence, active motion tracking, and biomechanical analysis significantly enhances the evaluation and effectiveness of fall prevention training, bridging the gap between education and objective outcome measurement. • Integrating AI and education improves fall prevention training, enhancing protective strategies and reducing injury risk among vulnerable populations. Topic Fall Prevention, Educational Intervention, Artificial Intelligence.