Awesome

LPS

Code for the paper ``Learning to Promote Saliency Detectors" accepted by CVPR 2018.

Usage

• first, install Pytorch
• modify the prior map, input and output directory in test_once.py, test_iterative.py, and train.py
• to download and use my trained model, run test_once.py or test_iterative.py (pretrained models can be downloaded here
• to train a new model, run train.py

to evaluate performance, run evafunc.py or call

evaluate('/path/to/images', 'path/to/ground/truth', 'path/to/save/results', 'name')

Citation

@inproceedings{zeng2018learning,
    author = {Yu Zeng, Huchuan Lu, Lihe Zhang, Mengyang Feng, and Ali Borji},
    title = {Learning to Promote Saliency Detectors},
    booktitle = {Proceedings of IEEE Conference on Computer Vision and Pattern Recognition,
    year = {2018}}