You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Losses can be chosen with the --losstype option in the arguments in train.py file. The valid options are currently: [Iou|Giou|Diou|Ciou|SmoothL1|Piou].
Fold-Structure
The fold structure as follow:
config/
config.py
init.py
data/
init.py
VOC.py
VOCdevkit/
model/
build_ssd.py
init.py
backbone/
neck/
head/
utils/
utils/
box/
detection/
loss/
init.py
tools/
train.py
eval.py
test.py
work_dir/
Environment
conda create --name probiou-ssd python=3.6
conda activate probiou-ssd
pip install -r requirements.txt
Datasets
PASCAL VOC:Download VOC2007, VOC2012 dataset, then put VOCdevkit in the data directory or run get_voc_dataset.sh in data folder
if you want to visual the box, you can add the command --visbox True(default False)
Cite our work
@article{Murrugarra_Llerena_2024,
title={Probabilistic Intersection-Over-Union for Training and Evaluation of Oriented Object Detectors},
volume={33},
ISSN={1941-0042},
url={https://dx.doi.org/10.1109/TIP.2023.3348697},
DOI={10.1109/tip.2023.3348697},
journal={IEEE Transactions on Image Processing},
publisher={Institute of Electrical and Electronics Engineers (IEEE)},
author={Murrugarra-Llerena, Jeffri and Kirsten, Lucas N. and Zeni, Luis Felipe and Jung, Claudio R.},
year={2024},
pages={671–681} }