This project will no longer be maintained by Intel.
Intel has ceased development and contributions including, but not limited to, maintenance, bug fixes, new releases, or updates, to this project.
Intel no longer accepts patches to this project.
If you have an ongoing need to use this project, are interested in independently developing it, or would like to maintain patches for the open source software community, please create your own fork of this project.
Brady Zhou, Philipp Krähenbühl, and Vladlen Koltun
Science Robotics, 4(30), 2019
Paper
Project
Code to accompany the paper.
git clone --recursive https://github.com/intel-isl/vision-for-action.git
We require building from source using the included submodule,
which includes hacks for getting labels for the walls and floors.
cd ViZDoom
cmake -DCMAKE_BUILD_TYPE=Release -DBUILD_PYTHON3=ON
make
Link ViZDoom and helpers to PYTHONPATH
.
export PYTHONPATH=$PYTHONPATH:$PWD/ViZDoom/bin/python3.6/pip_package
export PYTHONPATH=$PYTHONPATH:$PWD
This is supported on Windows machines only, and will take a little bit of effort.
You will need to install the following -
- a modified GameHook
- a modified GameHook GTA V Plugin
- a modified PyHookV
After successfully building all of these projects, you should have
dxgi.dll (renamed from gamehook.dll)
python.hk
server.hk
gta5.hk
pyhookv.pyd
Take all of these files and put them in the directory where GTAV.exe
lives.
Additionally, move the following files from the gta_v
directory to the same directory.
agents_privileged.py
constants.py
controller.py
controls.py
message_packer.py
pyhookv_utils.py
presets.py
scenarios.py
@article{Zhou2019DoesCV,
title={Does computer vision matter for action?},
author={Brady Zhou and Philipp Kr{\"a}henb{\"u}hl and Vladlen Koltun},
journal={Science Robotics},
volume={4},
number={30},
year={2019}
}