Python implementation of the Anomaly Detection Framework

A cleaner implementation of the Del Giorno, Bagnell, Hebert ECCV 2016 paper 'A Discriminative Framework for Anomaly Detection in Large Videos'

Building (C++ code)

Rather than cluttering the source directory with build files, run the following: cmake -Bbuild -H.; cd build/; make; cd ../

or with debug flags:

cmake -DCMAKE_BUILD_TYPE=Debug -Bbuild -H.; cd build/; make; cd ../

This will create a build in the build/ folder.

Using python virtualenv

Given that you'll need a set of Python packages to run, the best method to run this software is to install virtualenv, to mimic the same python system I developed in. Here are the steps:

Install virtualenv pip install virtualenv

Create the 'vanilla' virtual env virtualenv anomalyframework_virtualenv

Activate the virtual environment bash anomalyframework_virtualenv/bin/activate

Install the packages from requirements.txt pip install -r requirements.txt

When running code from this repository, you can then either activate the virtual environment each time (and use python, ipython as normal), or execute anomalyframework_virtualenv/bin/python to circumvent the activation, and call the executables directly.

Testing

run: python unit_tests/example_runscript.py