Generative adversarial networks to create synthetic motion capture datasets including subject and gait characteristics

• Published in: Journal of biomechanics Vol. 177; p. 112358
• Main Authors: Bicer, Metin, Phillips, Andrew T.M., Melis, Alessandro, McGregor, Alison H., Modenese, Luca
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.12.2024; Elsevier Limited
• Subjects: Adult; Biomechanical Phenomena; Conditional Generative Adversarial Networks; Data compression; Datasets; Deep Learning; Design; Female; Gait; Gait - physiology; Generative adversarial networks; Humans; Kinematics; Male; Middle Aged; Motion Capture; Movement; Neural networks; Neural Networks, Computer; Synthetic data; Synthetic Mocap Dataset; Walking; Walking - physiology; Gait; Conditional Generative Adversarial Networks; Synthetic Mocap Dataset

Resource-intensive motion capture (mocap) systems challenge predictive deep learning applications, requiring large and diverse datasets. We tackled this by modifying generative adversarial networks (GANs) into conditional GANs (cGANs) that can generate diverse mocap data, including 15 marker traject...