Pose Estimation Pipeline

[carlosj934]

Will look at either OpenPose or MediaPipe for this. I'll see what I decide as I go.

Regardless, this will be the goal:

Pose Estimation
Apply Pose Estimation Models: Use state-of-the-art pose estimation frameworks (like OpenPose, YOLOv8-Pose, ViTPose, or TensorFlow’s MoveNet) to detect and track body keypoints for each grappler in every frame. This converts video into structured pose data.

[carlosj934]

Example openpose pipeline, will need to look more into openpose: https://chatgpt.com/share/686b51d5-7db4-8000-a700-10193870261a

[carlosj934]

MediaPipe:

• Pose landmark detection
• Example python code for the landmark detection

[carlosj934]

Example of how a picture looks on mediapipe: