Pose estimation is a computer vision technique used to locate anatomical landmarks, such as the shoulders, hips, knees, and ankles, in an image or video. Connecting these landmarks creates a digital representation of body position, often called a skeleton.
In movement analysis, pose estimation can support joint tracking, angle visualization, and frame-by-frame observation without requiring physical markers to be attached to the body.
Factors That Affect Pose Estimation
- Lighting, image resolution, and motion blur
- Clothing and contrast with the background
- Body parts being hidden or overlapping
- Camera angle and distance
- Fast or unusual movement positions
Pose-estimation output is an estimate rather than a direct measurement of internal joint structures. Landmark positions should be visually checked, especially when they are used to interpret technique or compare recordings.