First make sure that you had cloned the original repository.
git submodule update --init
Then you should copy the necessary modified files and download pretrained networks. You could use the script in "360Tracking/code/" directory.
$360Tracking/code/
bash script_modify_TracKit.sh
After that, this guide is now exactly the same as tutorial on the original repository TracKit.
Run the installation script to install all the dependencies. You need to provide the conda install path (e.g. /home/user/anaconda3) and the name for the created conda environment (here TracKit).
$360Tracking/code/
cd TracKit/lib/tutorial
bash install.sh $conda_path TracKit
# get back to $360Tracking/code/TracKit
cd ..
cd ..
conda activate TracKit
python setup.py develop
Note: We have used only pytorch version, you could try TensorRT version as well (TracKit).
Get back to "360Tracking/code/" directory and run the following command.
$360Tracking/code/
# default Ocean
python TracKit/tracking/run_video_360.py --arch Ocean --resume TracKit/snapshot/OceanV19on.pth -v annotation/dataset-demo/demo-annotation/demo.mp4
# Ocean with equirectangular border improvement
python TracKit/tracking/run_video_360.py --arch Ocean --resume TracKit/snapshot/OceanV19on.pth -v annotation/dataset-demo/demo-annotation/demo.mp4 -border
# Pcean with NFOV improvement
python TracKit/tracking/run_video_360.py --arch Ocean --resume TracKit/snapshot/OceanV19on.pth -v annotation/dataset-demo/demo-annotation/demo.mp4 -nfov
# default SiamDW
python TracKit/tracking/run_video_360.py --arch SiamDW --resume TracKit/snapshot/siamdw_res22w.pth -v annotation/dataset-demo/demo-annotation/demo.mp4
# SiamDW with equirectangular border improvement
python TracKit/tracking/run_video_360.py --arch SiamDW --resume TracKit/snapshot/siamdw_res22w.pth -v annotation/dataset-demo/demo-annotation/demo.mp4 -border
# SiamDW with NFOV improvement
python TracKit/tracking/run_video_360.py --arch SiamDW --resume TracKit/snapshot/siamdw_res22w.pth -v annotation/dataset-demo/demo-annotation/demo.mp4 -nfov