HasMotion works with partners worldwide to ensure that our users have easy access to the latest developments in motion capture. One of those partners is Theia, the company behind the Theia3D markerless motion capture solution. This solution has been extensively validated and adopted worldwide, including in the sport of baseball.
Advances in markerless motion capture over the past decade have changed how we measure human movement in the real world. Markerless technologies like Theia3D allow biomechanists to collect scalable, reproducible measurements of people doing many different actions. They also help biomechanics get outside the lab, broadening collection possibilities. And recent changes to Theia3D’s output model is broadening what’s possible.
Biomechanics, Theia, and Baseball
Baseball has often been a proving ground for motion capture technologies applied to sports. Two key players, the pitcher and batter, stand in the same spot and repeat the same actions all game long. And they keep this up over a 162-game regular season plus playoffs. Now, inline with Theia’s newly tracked bat and ball segments, Visual3D’s automatic Theia model now includes both segments by default.
Theia advertises these newly tracked segments as helping users to avoid “stitching together data from multiple systems.” From our perspective this simplifies the analysis of markerless data combining both the human body and sports equipment. This provides biomechanists with a holistic view and greatly simplifies data management for analysis.
Visual3D’s automatic model leverages the C3D format’s support for multiple subjects. This makes it easy for users to work with different bats and balls, e.g., weighted balls, within a single session. It also helps users integrate this new motion capture feature into their established motion capture analysis routines. Working with this data in Sift allows users to efficiently scale their analysis across multiple games or to set up event-based triggers to quickly turnaround analysis.
Consistent Updates in Visual3D
These model changes are hardly unique, as Theia has pushed markerless pose estimation forward over the past 18 months. This notably included Theia’s Apollo and Axiom releases, with Apollo including changes that improved pose estimation across different activities.
Visual3D incorporated these model changes so that Theia3D users continue to enjoy seamless analysis of markerless data. In particular, Theia Apollo provided users with:
- A new pelvis definition, which mimics the Visual3D Composite pelvis for marker-based motion capture;
- Three different models: Theia3D’s default model, a Full-Body model, and a Separate Arms and Head model; and
- Estimates of each segment’s moment of inertia.
Full details of Visual3D’s automatic Theia model, as well as the different segment types used to build the model, are available on our wiki documentation site.
Looking Forward
Biomechanics has seen a shift in the last few years towards larger data sets. This shift has been fuelled in no small part by modern markerless motion capture technologies like Theia3D. These technologies allow researchers to collect motion capture data in streetwear, of world-record efforts in competition, and outdoors on the side of mountains. And the ease of collecting makes it possible to collect a lot, across many sites or for ongoing studies.
Visual3D and Sift continue to adapt to the latest features in Theia3D, making these advances accessible to our mutual users and supporting high quality biomechanical analysis.