Benchmarking dataset for markerless motion capture analysis

This dataset contains data from a motion analysis experiment using a markerless and a marker-based system. Its purpose is to benchmark whole-body markerless motion analysis methods. A GitHub repository (https://github.com/lbmc-lyon/Benchmarking_markerless) is associated with this dataset, where one may find and upload kinematics obtained from video data. The data consisted of recordings with both systems of five different tasks performed by two participants. The tasks cover various activities and challenges for both systems (gait, sit-to-stand-to-sit transfers, manual box handling, challenging motions with high feet, and couple dancing). The dataset includes: "marker" data (3D trajectories of 48 skin reflective markers placed on specific anatomical landmarks) provided as .c3d files; RGB video from 9 calibrated and synchronized cameras provided as .avi files; Utility files such as camera calibration files.