3D Markerless Motion Capture

In 2024 TCPR took delivery of a general state-of-the-art 3D motion capture system. Since then the system has been optimized to specifically address tennis mechanics. This new system represents a major upgrade to previous systems we have operated. The previous systems required that players be fit with reflective or infrared anatomical markers. These markers significantly interfered with performance of the strokes. The advancement of the new system is that scientific grade 3D capture is accomplished without anatomical markers of any kind. Players simply hit in their normal attire – this allows unhindered 3D capture in any situation including match play

Simi Reality Motion Systems based in Munich, Germany has been at the forefront of the markerless revolution from the beginning. TCPR and Simi have forged a developmental relationship with the primary goal of creating a turnkey application dedicated to tennis player development.

3D motion capture has traditionally been too expensive and complex to implement in training environments. Simi will contribute experience in motion capture technology and combine it with TCPR experience in coaching and tennis biomechanics. The result will be a research grade sport science application built for training facilities and indispensable for coaches and players.

An outdoor laboratory has been constructed on a court at the Rick Macci Tennis Center (Court 16). It consists of 8 synchronized high speed cameras operated by a custom computer system at court-side. The court play area is calibrated prior to motion capture in a comprehensive procedure that allows the cameras to work together to reconstruct a 3D silhouette (shape) of the body during strokes. Once calibrated, the cameras record the stroke(s) as the first step of the process.

Once a stroke is recorded, the 3D shape is computed. This process is referred to as “tracking”. The process combines computer vision methods and extensive use of AI learning algorithms. The body, racquet, and ball are tracked one video frame at time for each camera. The information from all cameras is combined to compute the 3D silhouette. The video shows the actual tracking of a portion of a forehand. The colored markings (green) on the player show where the computer is locating body landmarks for 3D reconstruction.

The 3D shape produced by the tracking procedure may now be defined by its skeletal structure. Further use of AI driven algorithms, combined with reference data from biomechanical and anatomical research, guide fitting of a skeletal infrastructure to the 3D shape.

The skeletal structure represents complete numerical expression (mapping) of body anatomy during the stroke. This may be represented as a “stick figure” to define the 3D location of the body joints. The mapping data may also be used for Biomechanical computations that illuminate the true nature of the stroke. The stick figure representation of the skeletal structure is shown in the video.