This tool uses MediaPipe's deep learning pose estimation model to annotate images of humans with body landmarks.
From these landmarks, features are created for a machine learning model.
Using an XGBoost model trained on aerial straps poses, the pose is then classified.
Additionally, a visualization chart displays the confidence level for each pose classification.
For more details, code, and information, check out the project repository on GitHub.
Upload your image to proceed with pose annotation and classification.