This package has the source code for the paper "Kernelized Subspace Ranking for Saliency Detection" (ECCV16).

The paper can be found in Baidu Drive or Google Drive.

1. Install prerequsites for Caffe (see: Caffe installation instructions)
2. Compile the ./sds_eccv2014-master/extern/caffe-mastersubmodule.
3. Compile the ./gop_1.3 submodule.

Download pretrained SDS model from Baidu Drive or Google Drive. Then put it into the ./sds_eccv2014-master folder.

1. Train: run runme_train.m to generate trained model in the ./trained_model folder.
2. Test: run runme_test.m to generate saliency maps in the ./saliency_map folder.

Download trained models from Baidu Drive or Google Drive. Then put it into the ./trained_model folder.

Download saliency map from Baidu Drive or Google Drive.

If you find this work useful in your research, please consider citing:

 @inproceedings{wangeccv16,
    Author={Tiantian Wang and Lihe Zhang and Huchuan Lu and Chong Sun and Jinqing Qi},
    Title={Kernelized Subspace Ranking for Saliency Detection},
    Booktitle={European Conference on Computer Vision (ECCV)},
    Year={2016}
 }

[email protected]