Main Page → Motive Documentation → Assets → Skeleton Tracking
In Motive, skeleton assets are used for tracking human motions. These assets auto-label specific sets of markers attached to human subjects, or actors, and create skeletal models. Unlike rigid body assets, skeleton assets require additional calculations to correctly identify and label 3D reconstructed markers on multiple semi-rigid body segments. In order to accomplish this, Motive uses pre-defined skeleton markerset templates, which is a collection of marker labels and their specific positions on a subject. According to the selected markerset, retroreflective markers must be placed on pre-designated locations of the body. This page details instructions on how to create and use skeleton assets in Motive.
When it comes to tracking human movements, a proper marker placement becomes especially important. Motive utilizes pre-programmed skeleton markersets, and each marker is used to indicate anatomical landmarks when modeling the skeleton. Thus, all of the markers must be placed at their appropriate locations. If any of markers are misplaced, the skeleton asset may not be created, and even if it is created, bad marker placements may lead to labeling problems. Thus, taking extra care in placing the markers on intended locations is very important and can save time in post-processing of the data.
Attaching markers directly onto a person’s skin can be difficult because of hairs, oils, and moistures from sweat. Plus, dynamic human motions tend to move the markers during capture, so use appropriate skin adhesives for securing marker bases onto the skin. Alternatively, mocap suits allow velcro marker bases to be used.
Open Builder pane and go to the skeleton creation feature. Select the markerset you desire to use from the drop-down menu. A total number of required markers for each skeleton is indicated in the parenthesis after each markerset name, and corresponding marker locations are displayed over an avatar that shows up in the Builder pane. Instruct the subject to strike a calibration pose (T-pose or A-pose) and carefully follow the figure and place retroreflective markers at corresponding locations of the actor or the subject.
All markers need to be placed at respective anatomical locations of a selected skeleton as shown in the Builder pane. Skeleton markers can be divided into two categories: markers that are placed along joint axes (joint markers) and markers that are placed on body segments (segment markers).
When using the biomechanics markersets, markers must be placed precisely with extra care because these placements directly relate to coordinate system definition of each respective segment; thus, affecting the resulting biomechanical analysis. The markers need to be placed on the skin for direct representation of the subject’s movement. Mocap suits are not suitable for biomechanic applications. While the basic marker placement must follow the avatar in the Builder pane, additional details on the accurate placements can be found on the following page: Biomechanics Markersets.
The magenta markers indicate the segment markers that can be placed at a slightly different position within the same segment.
In Edit Mode
Virtual Reality Markersets
By configuring Skeleton Properties, you can modify the display settings as well as skeleton creation pose settings for skeleton assets. For newly created skeletons, default skeleton creation properties are configured under the Application Settings pane. Properties of existing, or recorded, skeleton assets are configured under the Properties pane while the respective skeletons are selected in Motive.
A proper calibration posture is necessary because the pose of the created skeleton will be calibrated from it. Read through the following explanations on proper T-poses and A-poses.
The T-pose is commonly used as the reference pose in 3D animation to bind two characters or assets together. Motive uses this pose when creating skeletons. A proper T-pose requires straight posture with back straight and head looking directly forward. Both arms are stretched to each side, forming a “T” shape. Both arms and legs must be straight, and both feet need to be aligned parallel to each other.
The A-pose is another type of calibration pose that is used to create skeletons. Set the Skeleton Create Pose setting to the A-pose you wish to calibrate with. This pose is especially beneficial for subjects who have restrictions in lifting the arm. Unlike the T-pose, arms are abducted at approximately 40 degrees from the midline of the body, creating an A-shape. There are three different types of A-pose: Palms down, palms forward, and elbows bent. ↑
Calibration markers exists only in the biomechanics markersets
Many skeleton markersets do not have medial markers because they can easily collide with other body parts or interfere with the range of motion, all of which increase the chance of marker occlusions.
However, medial markers are beneficial for precisely locating joint axes by associating two markers on the medial and lateral side of a joint. For this reason, some biomechanics markersets use medial markers as calibration markers. Calibration markers are used only when creating skeletons but removed afterward for the actual capture. These calibration markers are highlighted in red from the 3D view when a skeleton is first created.
After creating a skeleton from the Builder pane, calibration markers need to be removed. First, detach the calibration markers from the subject. Then, in Motive, right-click on the skeleton in the perspective view to access the context menu and click Skeleton → Remove Calibration Markers. Check the assigned marker positions to make sure that the skeleton no longer expects markers in the corresponding medial positions.
Existing skeleton assets can be recalibrated using the existing skeleton information. Basically, the recalibration recreates the selected skeleton using the same skeleton markerset. This feature recalibrates the skeleton asset and refreshes expected marker locations on the assets.
To recalibrate skeletons, select all of the associated skeleton markers from the perspective view and click Recalibrate From Markers which can be found in the skeleton context menu from either the Assets pane or the Perspective View pane. When using this feature, select a skeleton and the markers that are related to the corresponding asset.
Skeleton recalibration do not work with skeleton templates with added markers.
Skeleton marker colors and marker sticks can be viewed in the perspective view pane. They provide color schemes for clearer identification of skeleton segments and individual marker labels from the perspective viewport. To make them visible, enable the Marker Sticks and Marker Colors under the Application Settings or under the visual aids in the perspective view pane. Default color scheme is assigned when creating a skeleton markerset. To modify them, export and edit the skeleton template XML file, where the custom marker labels can also be assigned.
The marker colors and sticks are featured only in Motive 1.10 and above, and skeletons created using Motive versions before 1.10 will not include the colors and sticks. For the Takes recorded before 1.10, the skeleton assets will need to be updated from the Assets pane by right-clicking onto an asset and selecting Update Markers. The Update Markers feature will apply the default XML template to skeleton skeleton assets.
Skeleton markersets can be modified slightly by adding or removing markers to or from the template. Follow the below steps for adding/removing markers.
Note that modifying, especially removing, skeleton markers is not recommended since changes to default templates may negatively affect the skeleton tracking when done incorrectly. Removing too many markers may result in poor skeleton reconstructions while adding too many markers may lead to labeling swaps. If any modification is necessary, try to keep the changes minimal.
When adding, or removing, markers in the Edit mode, the Take needs to be auto-labeled again to re-label the skeleton markers.
Assets can be exported as either MOTIVE (.motive) or CSV (.csv) files, and they can be re-imported into Motive when needed. Exported files are text-readable files that can contain various configuration settings in Motive; including the asset definitions.
When the asset definition(s) is exported, it stores marker arrangements calibrated in each asset, and they can be imported into different takes without creating a new one in Motive. Note that these files specifically store the spatial relationship of each marker, and therefore, only the identical marker arrangements will be recognized and defined with the imported asset.
To export the assets, go to Files tab → Export Assets to export all of the assets in the Live-mode or in the current TAK file. You can also use Files tab → Export Profile to export other software settings including the assets.
Important Update Note
TRA/SKL files can still be imported into Motive, but they will be deprecated from the next release. This functionality will be replaced with the motive profile functionality. Starting from Motive 2.1, you can export just the Asset definitions into a Motive profile) and re-import them whenever necessary. In Motive 2.1, you will no longer be able to export out TRA/SKL files.
There are two ways of obtaining skeleton joint angles. Rough representations of joint angles can be obtained directly from Motive, but the most accurate representations of joint angles can be obtained by pipelining the tracking data into a third-party biomechanics analysis and visualization software (e.g. Visual3D or The MotionMonitor).
For biomechanics applications, joint angles must be computed accurately using respective skeleton model solve, which can be accomplished by using a biomechanical analysis software. Export C3D files or stream tracking data from Motive and import them into an analysis software for further calculation. From the analysis, various biomechanics metrics, including the joint angles can be obtained.
Joint angles generated and exported from Motive are intended for basic visualization purposes only and should not be used for any type of biomechanical or clinical analysis. A rough representation of joint angles can be obtained by either exporting or streaming the skeleton rigid body tracking data. When exporting the tracking data into CSV, set the Coordinate Space setting to Local to obtain bone segment position and orientation values in respect to its parental segment, roughly representing the joint angles by comparing two hierarchical coordinate systems. When streaming the data, set Local Rigid Bodies to true in the streaming settings to get relative joint angles. But, again.
Each skeleton asset has its marker templates stored in an XML file. By exporting, customizing, and importing the skeleton XML files, a skeleton markerset can be modified. Specifically, customizing the XML files will allow you to modify skeleton marker labels, marker colors, and marker sticks within a skeleton asset. For detailed instructions on modifying skeleton XML files, read through Skeleton Tracking: Marker Name XML Files page.