Skeletal Tracking
Also known as: stick-figure tracking, joint tracking
Skeletal tracking is software that finds a person's joints in video and links them into a stick figure. It lets one phone video estimate body positions.
Skeletal tracking is a computer-vision process that scans each video frame. It identifies where the major joints (shoulders, hips, elbows, wrists, knees, ankles) most likely are. It then connects them into a simplified moving stick figure, overlaid on the original footage. This stick figure is the raw material for every later measurement, such as rotation angles, posture changes, and weight distribution estimates.
Tracking quality depends heavily on video conditions. A clear, well-lit video with the whole body visible and minimal background clutter produces a much more reliable skeleton. It beats a dim, cropped, or busy video, where loose clothing or an unusual camera angle can confuse the joint-detection model. This is why camera angle guidance and lighting matter so much for any video-based swing analysis tool.
Skeletal tracking is the same underlying category of technology used in sports broadcasting overlays, fitness apps, and some smartphone health features. Golf swing analysis is one particular application of a much broader computer-vision capability.
Example
Software processing a swing video draws a simplified stick figure over the golfer's body, tracking how the shoulder and hip joints rotate frame by frame through the backswing and downswing.
Common mistakes
- Filming in low light, baggy clothing, or with the body partially out of frame, all of which degrade how reliably the joints can be identified and tracked.
In SwingVantage Motion Lab
Skeletal tracking is how Motion Lab reads a swing: it finds one athlete and follows their joints frame by frame. After the analysis, a capture check reports whether your whole body stayed in frame and how reliably the joints were tracked. Dim light, blur or a cropped body lowers the confidence shown beside each read. The club is not part of the skeleton, so it is not tracked.
Related terms
- Pose EstimationPose estimation identifies a person's body position, joint by joint, from a 2D video frame. It is the computer-vision technique behind markerless swing apps.
- Motion CaptureMotion capture tracks a golfer's body in 3D to build a precise digital skeleton for swing analysis. It traditionally uses reflective markers and several cameras.
- 3D Swing Analysis3D swing analysis reconstructs the swing from motion capture or several cameras. Angles like hip rotation are then measured, not estimated from a flat image.
- Analysis Confidence LevelAnalysis confidence level is a stated measure of how reliable a video-derived swing observation is. It keeps a rough estimate from passing as a certain fact.
Related guides & benchmarks
Put this into your swing
SwingVantage can spot this in your own swing — free to start.
See a sample Golf report first