Skip to content

Latest commit

 

History

History
28 lines (25 loc) · 725 Bytes

README.md

File metadata and controls

28 lines (25 loc) · 725 Bytes

pose_mimic

This project is build as a ROS package and uses YOLO-v7 models for human pose detection and instance segmentation. The algorithm is based on this paper

Requirements

  • ZED camera is connected
  • ZED ROS Wrapper is running
  • Model weights are downloaded. Run the following command to download the weights
cd weights
bash download_weights.sh

Testing

Source the package

source ~/catkin_ws/devel/setup.bash

Run the human pose detection node

cd scripts/yolov7_pose
python3 pose_estimate.py

Run the instance segmentation node

cd scripts/yolov7_segmentation
python3 segmentation.py