SPA_CVPR2021

Xingjia Pan, Yingguo Gao, Zhiwen Lin, Fan Tang, Weiming Dong, Haolei Yuan, Feiyue Huang, Changsheng Xu

The official implementaion of SPA_CVPR2021 paper "Unveiling the Potential of Structure-preserving for Weakly Supervised Object Localization"

Setup

Create an Anaconda enrironment with python>=3.6

Clone this repo:

SPA_ROOT=/path/to/SPA
git clone https://github.com/Panxjia/SPA_CVPR2021.git $SPA_ROOT
cd $SPA_ROOT
pip install -r requirements.txt
pip install torch==1.1.0 torchvision==0.2.1

Prepare the ILSVRC2012 dataset
1. Download training and validation datasets, and put them at $SPA_ROOT/data/ILSVRC
2. Download namelist and put it at $SPA_ROOT/data/ILSVRC
  The dataset is organised as:
```
ILSVRC-
     -- Train
         -- n01440764
         -- ...
     -- val
         -- ILSVRC2012_val_00000001.JPEG
         -- ...
     -- list
```
Download the pretrained models VGG16 and Inception V3, and place them at $SPA_ROOT/pretrained_models/

Train and test

Train
```
cd scripts
bash train_spa_ilsvrc.sh
```
Test

Download our models VGG16 and place it at $SPA_ROOT/snapshots/vgg_16_baseline_ilsvrc_20e_10_15d
```
cd scripts
bash val_spa_ilsvrc.sh
```

Citation

If you find this project useful for your research, please use the following BibTeX entry.

@article{pan2021unveiling,
  title={Unveiling the Potential of Structure Preserving for Weakly Supervised Object Localization},
  author={Xingjia Pan and Yingguo Gao and Zhiwen Lin and Fan Tang and Weiming Dong and Haolei Yuan and Feiyue Huang and Changsheng Xu},
  booktitle={CVPR},
  pages={1--8},
  year={2021}
}

Contacts

If you have any questions about our work, please do not hesitate to contact us by emails.
Xingjia Pan: xjia.pan@gmail.com
Fan Tang: tfan.108@gmail.com

Panxjia / SPA_CVPR2021