Martial Arts & Motion Capture

Calibration

Before any motion capture session can begin, one must first calibrate the physical volume in which the motion capture actor will perform their act with the virtual volume. This is done by waving a wand with markers on it around the physical space. This creates data points for the optical camera system to accurately determine the position of each camera within the room. Once the software (Vicon Blade) has determined the position of each optical camera, the software must then be told the directions of the axis. Some 3D software packages differ when it comes to axis direction. Unreal Engine for example uses Z-Up coordinate system whereas Autodesk Maya defaults to Y-Up. In order to accurately coordinate the axis direction with the software package we may export to, what’s called an ErgoCal L-Frame must be placed at the origin point in the room. In this case, an ErgoCal L-Frame Y-Up was used since the motion capture data was being exported for Autodesk Maya.

Calibrating optical cameras using 240mm Wand.

Setting the origin using ErgoCal L-Frame Y-Up.

Marker Placement

In order to collect and capture the skeletal data from a motion capture actor, one must place tracking points on a subject in order to allow cameras to follow a certain point through space. In this case, 45 markers were placed on various points on my body.

Once the markers have been placed on the subject, what is called a ROM, also known as a Range of Motion must be performed by the subject. The ROM captures the subject from a T-pose position into a wide range of motions. The subject will rotate their arms, head, shoulders, elbows, shoulders, hips, legs, knees, and feet. Similar to the calibration stage, the range of motion allows the software to determine the separation distance between various points on a motion capture actor. Once the range of motion is performed, the tracking point data is calibrated to a template skeleton. This labels all of the tracking points as skeletal segments such as L_Hip (Left Hip), R_Elbow (Right Elbow) etc.

Performing Range of Motion (ROM).

Calibrated tracking points with template skeleton.

Subject Capture

The purpose of this study was to determine the best method to capture the details of a human body when doing motion capture for martial arts sequence in order to teach that art in a more modern 3D manner. For a Karate instructor or competition coach, being able to use a motion capture system to analyze a martial artists’ skeletal data is highly valuable because they are able to make suggestions for improved demonstration quality. Certain adjustments such as stance height, stance width, elbow separation from body, hand placement, kick height, punch / kick velocities, and much more can be made in post-production software such as Autodesk Maya 2016.

Demonstrating a “Back Stance”.

Demonstrating a “Horse Stance”.

Tracking Fingers

In martial arts, there are many instances where one will clench their fist or open their hand to perform either blocking or attacking maneuvers. The problem with tracking fingers for martial arts was that the markers were often occluded by the rest of the fingers when making a fist for punching. The markers on the fingertips would bend into the center of the fist; therefore the cameras were unable to track their positions.

15 markers placed on each hand.

Finger marker data in Vicon Blade.