This project is considered the first step of generating dances based on music genre & rhythm. This part is used to extract people patterns from the video using SSD (Single Shot Detector). then detect the body keypoints (joints) in order to represent the body skeleton. Caffe models trained on COCO database are used for both detectors. To download the movement caffe model: https://drive.google.com/file/d/15PROIfmF35Gpde8-UDyBQvDswBXQ8dGT/view?usp=sharing
Thanks to PyImageSearch & LearnOpenCV for their great tutorials