Large, diverse datasets will be essential to harness the power of data science in rehabilitation. We have created what is currently the world’s largest human movement dataset, which will continue to grow through use of our Center’s software tools. This will enable a new scale of research, and the diversity of the dataset will enable our methods to better represent the diverse demographics of people with movement limitations.
Datasets
These datasets, derived from many past and ongoing studies, include lab-based motion capture data that has been carefully curated and reviewed by our team to ensure that the data meets verification and validation guidelines. The current release of the data includes 300 hours of high-quality, lab-based experimental movement data (i.e., marker data from optical motion capture, ground reaction forces, and/or IMU data) and demographic and diagnosis information from over 1,300 individuals, along with the corresponding models and 3D simulations of movement (i.e., joint kinematics and kinetics).
| Dataset | Number of Participants | Number of Trials | Data Description |
| AddBiomechanics Dataset | 461 | 8,139 (71 Hours) | – Movements: Walking, jumping, cutting, squatting, sit-to-stand, stair traversal, and many others. – Mocap, 80% with GRFs |
| Marker Enhancer Dataset | 1,071 | 23,636 (221 Hours) | – Movements: Walking, jumping, cutting, squatting, sit-to-stand, stair traversal, dancing, reaching, and many others. – Mocap, video keypoints |
| Kinematics and Timed Function Tests of Facioscapulohumeral Muscular Dystrophy and Myotonic Dystrophy | 129 | – Movements: Running, walking, timed up and go, sit to stand, Brooke Upper Extremity Scale, and more. – OpenCap keypoints |
Dataset Growth
The FAIR Center datasets are automatically growing via our software tool, AddBiomechanics, that ingests, harmonizes, processes, and evaluates the accuracy of new datasets as they become available. AddBiomechanics is a web-based tool to automatically process optical motion capture and ground reaction force data to scale musculoskeletal models and compute kinematics and kinetics.
We will also enhance our OpenCap software to similarly add the data shared through it to the FAIR Center datasets. OpenCap is an open-source platform for computing both the kinematics (i.e., motion) and dynamics (i.e., forces) of human movement using videos captured from two or more smartphones.