Stay organized with collections
Save and categorize content based on your preferences.
The MediaPipe Holistic Landmarker task lets you combine components of the
pose, face, and
hand landmarkers to create a complete landmarker for
the human body. You can use this task to analyze full-body gestures, poses, and
actions. This task uses a machine learning (ML) model on a continuous stream of
images. The task outputs a total of 543 landmarks (33 pose landmarks, 468 face
landmarks, and 21 hand landmarks per hand) in real-time.
An upgraded version of this MediaPipe Solution is coming soon! The MediaPipe Legacy
Solution for this task is available on
GitHub.