This repository contains Python scripts for extracting human poses from videos using the YOLO (You Only Look Once) object detection model. The project consists of two main scripts: pose.py and pose_extract.py.
This script uses the YOLO model for detecting human figures in a video and extracts their poses. It reads a video file, processes each frame to detect human figures, and then extracts their pose keypoints.
os cv2 (OpenCV) numpy pandas matplotlib ultralytics (YOLO)
To use this script, ensure you have a video file in the specified directory and run the script. The output will be the video frames with detected human poses.
This script contains functions to assist in the extraction and processing of pose keypoints from the detected human figures.
The script includes definitions for various key points in the human body such as the nose, eyes, shoulders, elbows, hands, hips, knees, and feet.
This is a helper script used by pose.py and is not intended to be run independently.
Provide an example of how to run the script, such as:
python pose.py