Gait2Hip-60: A Multi-Cadence Gait Dynamics Dataset

Gait2Hip-60 is a multi-cadence gait dynamics dataset from 60 healthy subjects. The dataset provides trial-level OpenSim-derived gait biomechanics data, including inverse kinematics, inverse dynamics, and static-optimization-derived muscle force outputs, together with NPZ files for machine learning and deep learning applications. Code: https://github.com/ygsiete7/Gait2Hip-60 Paper: https://doi.org/10.48550/arXiv.2605.30374 Potential applications Gait2Hip-60 can be used for research on gait biomechanics, musculoskeletal modeling, hip dynamics analysis, machine learning-based biomechanical estimation, and time-series modeling. It can also support the development and evaluation of sequence models such as RNN, LSTM, Transformer, and other deep learning architectures.