Digital motion capture (“mocap”) photography has a growing list of medical applications. Healthcare professionals use motion capture to assess injuries, to create and refine treatment plans, and to observe rehabilitation progress. Sports trainers use motion capture video to help athletes perform at their peak. Most current motion capture technology requires that the subjects wear a specially marked suit or reflective tape markers on their body. Mocap suits are most common with motion capture for 3D animation, but medical mocap applications require precisely attached markers. Attaching the markers is a time-consuming process by specially trained personnel. Also, because precise placement is crucial, mocap subjects need to be careful not to move the markers, which can affect their natural movement.
Engineers at Nanyang Technological University, Singapore (NTU Singapore) and the Rehabilitation Research Institute of Singapore (RRIS) have developed a methodology for accurate marker-less motion capture. The technology, called Precise Marker-less, employs multiple cameras to capture and analyze human motion without the aid of traditional markers.
Precise Marker-less uses machine leaning to analyze movement capture from two to four cameras. In training the system, the team used more than 10 million subject movement images. The team published a research report on the marker-less motion capture system. As of March 2022, according to an NTU news release, this data amounts to 16 TB of data. In the reported testing, the system produced 3D bone landmark locations with an accuracy of 10-15 millimeters: roughly 1/2 inch.
RRIS Associate Professor Ang Wei, said, “…we harnessed the power of machine learning to train a computation model to predict the location of markers on video footage of a moving subject.”
The NTU/RRIS model is suitable for many current medical mocap applications, aside from some that require greater accuracy such as surgery. According to NTU, the savings in time and money with rehab or training applications is significant, reducing consultation and data analysis tasks.
