Keypoint annotations

Keypoint annotation is the process of labeling specific points or landmarks on an object—typically on human bodies, animals, or mechanical parts—to track movement, shape, or posture in images or videos.


🔹 Key Uses:

  • Human pose estimation (e.g., tagging joints like elbows, knees, wrists)
  • Facial landmark detection (eyes, nose, mouth corners)
  • Gesture and activity recognition
  • Medical imaging (skeletal structure analysis)
  • Autonomous driving (identifying key parts of pedestrians or cyclists)

🔹 How It Works:

Used to train models that analyze motion, expression, or structure.

Annotators mark specific points on the object (e.g., 17 points for a full human pose).

The positions are stored as x, y (and sometimes z) coordinates.