This is the research code for the paper:
Dong Li, Jia-Bin Huang, Yali Li, Shengjin Wang, and Ming-Hsuan Yang. "Weakly Supervised Object Localization with Progressive Domain Adaptation" In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016
If you find the code and pre-trained models useful in your research, please consider citing:
@inproceedings{Huang-CVPR-2016,
author = {Dong, Li and Huang, Jia-Bin and Li, Yali and Wang, Shengjin and Yang, Ming-Hsuan},
title = {Weakly Supervised Object Localization with Progressive Domain Adaptation},
booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition)},
year = {2015},
volume = {},
number = {},
pages = {}
}
- MATLAB (tested with R2014a on 64-bit Linux)
- Caffe
-
Download and unzip the project code.
-
Install Caffe. Please follow the Caffe installation instructions to install dependencies and then compile Caffe:
# We call the root directory of the project code `WSL_ROOT`. cd $WSL_ROOT/caffe-wsl make all -j8 make pycaffe make matcaffe
-
Download the PASCAL VOC 2007 dataset. Extract all the tars into one directory named
VOCdevkit
. It should have this basic structure:$VOCdevkit/ # development kit $VOCdevkit/VOCcode/ # VOC utility code $VOCdevkit/VOC2007 # image sets, annotations, etc. # ... and several other directories ... # Then create symlinks for the dataset: cd $WSL_ROOT/data ln -s $VOCdevkit VOCdevkit2007
-
Download the pre-trained ImageNet model and put it into
$WSL_ROOT/data/imagenet_models
. -
Download the pre-computed EdgeBox proposals and put them into
$WSL_ROOT/data/edgebox_data
. -
Install the project.
cd $WSL_ROOT # Start MATLAB matlab >> startup
You will need about 150GB of disk space free for the feature cache (which is stored in $WSL_ROOT/cache
by default. The final adapted model will be stored in $WSL_ROOT/output/default/voc_2007_trainval
.
-
Classification adaptation.
>> prepare_for_cls_adapt cd $WSL_ROOT sh cls_adapt.sh
-
Class-specific proposal mining.
>> maskout
-
MIL for confident proposal mining.
>> mil
-
Detection adaptation.
>> prepare_for_det_adapt cd $WSL_ROOT sh det_adapt.sh
-
Evaluation.
cd $WSL_ROOT sh test.sh