Skip to content

Latest commit

 

History

History
30 lines (19 loc) · 828 Bytes

README.md

File metadata and controls

30 lines (19 loc) · 828 Bytes

Pedestrian tracking

Simple tracking all pedestrian in video.

Theory about Convolutional Neural Network on my channel (CZ):

Video results here:

1) Conda instalation:

  • conda create -y --name pedestrian_tracking python==3.7
  • conda install -c conda-forge -y --name pedestrian_tracking --file requirements.txt

2) Download YOLOv3 weights to root project directory:

3) Demo command:

  • python3 demo.py

4) Run example:

  • save output to video file: python3 main.py --video_path videos/soccer_01.mp4 --output_video_path "out.mp4"

  • process video without result saving, only view: python3 main.py --video_path videos/soccer_01.mp4