Optitrack : Skeletons & Rigid Bodies overview

Tutorial

 How to use skeletons and rigid bodies : setup process, characteristics and differences

Dates
  • Creation: 09/28/2020
  • Update: 09/29/2020
Motive gives you the choice : you can either track an asset/ a solid (eg : a drone) or a human body. Depending on what you want to study you will have to choose between two tracking modes : rigid bodies or skeleton.
Each mode has its own caracteristics and tracking parameters, which we will see more in details in the corresponding tutorial. However, here is an overview of solid bodies and skeletons, which will enable you to apprehend them more easily.

Preparing the setup :

An essential element for both rigid bodies and skeletons is your markerset. Indeed, those markers are essential for the tracking process. Thus, you will have to provide them a special attention, and all the accuracy and precision of your work lies in a good preparation and pre-processing of your markerset.

About markers : You can use passive retro-reflective markers or active LED markers on the asset you want to track. They will track position and orientation with 6 degrees of freedom.

Get your markers and attach them on the subject/ object : try to secure them as much as possible, to protect them from any movment. Minimize extraneous reflections by covering shiny surfaces with non-reflective tape.

Attaching markers directly onto a human can be difficult : skin can be oily, hairs can get in the way, even moistures from sweat and dynamic motions tend to move the markers during the capture. It is important to use appropriate skin adhesives for securing marker bases onto the skin. You can also use suits that allow velcro marker bases to be attached.

Markers placement :

While placing the markers, try to avoid confusion and overlapping between them. Hence, opt for asymmetry : place your joint markers first on anatomical landmarks, then, place segment markers in asymmetrical positions for similar rigid bodies and skeletal segments. It will enable a clear distinction between two similar arrangements.

For example, with a square marker arrangement, you will have troubles with the orientation and frequent mislabels may occur throughout the capture. Thus, always try to place the less critical markers in asymmetrical arrangements. In general, avoid symmetrical shapes such as squares, isoceles or equilateral triangles.

Marker placement is a key step for motion capture data because each marker is used as an indicator for both position and orientation. Tracking results are calibrated and recorded in real time on Motive. Then, recorded information are used to recognize the markers and auto-label the asset. Asymmetry is vital for the auto-labelling process in order to avoid the congruency between markers, especially when you have multiple subjects with a similar markerset shape.

Tip : for more accurate 3D orientation data, you should spread markers as far as you can within the same rigid body. Hence, any deviation in the orientation, as slight as it could be, will be reflected from small changes in the position.

The more markers you will have, the better you will avoid marker occlusion : while recording, if any of the markers is occluded, Motive will be able to refer to other visible markers to compute the position and the orientation of the missing data.

However, having too many markers is not recommended ! It may increase the likelihood of overlapping and label-swaps during the capture.

Tip : For rigid bodies for example, 4 to 12 markers are recommended.

Modifying the markerset :

For tracking human movements, Motive uses pre-programmed skeleton markersets. Each marker aims to indicate a specific anatomical landmark when modeling the skeleton.

Be careful : you have to respect the proper location of a given marker. If a marker is misplaced, the skeleton asset may not be created and even if it is, bad marker placements lead to labeling problems.

When it comes to already existing shapes, keep in mind the fact that you can always modify them. However, you have to pay extra-attention to the settings :

  • Select all the markers from the body you want to modify (use the Assets pane or directly from the Perspective View pane)
  • CTRL+LEFT click the marker(s) you want to add/remove
  • LEFT-CLICK on the Perspective View pane to open the rigid body menu
  • Choose the Add/Remove selected markers to/from rigid bodies
  • RIGHT-CLICK on the rigid body and select Reset pivot to relocate the pivot point to the geometric new center

/!\ About the pivot point /!\

The pivot point is an essential data for the tracking process with rigid bodies ! Indeed, it is used to define both position and orientation. When a rigid body is created, Motive automatically sets the pivot point at its geometric center by default. Thus, its orientation axis will be aligned with the global coordinate axis. Sometimes, when adding or removing a marker from a rigid body its global shape might be affected : resetting the pivot center (if needed) is very important for having accurate 2D/ 3D-data and for post-processing calculation.

However, in some specific cases, the desired pivot point location is not the center of the rigid body. Hence, it can not be computed automatically by Motive. In this case, you will have to assign it manually :

  • Attach a marker on the desired pivot location
  • Set the pivot point to the marker
  • Apply a translation onto the body for precise adjustments