Activity Classification using MHI

Author: Richard Lucas

Writeup and Blog Post can be found here http://blog.richardalucas.com/2017/12/03/Activity-Classification-using-MHI-Techniques/

Introduction

Getting Started

The following dependencies will need to be installed:

• OpenCV for Python 2.x
• Numpy
• SciKit Learn
• Matplotlib

Setup

This creates the necessary folders and downloads all the needed files: $ python setup.py

It typically takes about 30 minutes to get everything.

How to use

Run the experiment

To run the experiment and recreate the figures in the report:
$ python experiment.py

It will output all figures in png format into the output/ directory It will output the multiple activity videos to the output_videos directory

Classify a video

If you just want to classify and create a video run: $ python ac.py [video_file] [output_name]

It will output an mp4 classified video in the output_videos/ directory

Example:
$ python ac.py Test1.mp4 test1

==> output_videos/classified_test1.mp4

Links

• Multiple Activity Videos:
  • Handclapping & Boxing Video: https://youtu.be/dEtfgAYwCg4
  • Multitude activity Video: https://youtu.be/x_hFKZzc0jw