Salient Object Detection in the Deep Learning Era: An In-Depth Survey

===========================================================================

Wenguan Wang, Qiuxia Lai, Huazhu Fu, Jianbing Shen, Haibin Ling

===========================================================================

It is very welcome to send me your saliency maps if your work is published in top-level conference.

If I miss your work, please let me know and I'll add it.

===========================================================================

Google Disk: https://drive.google.com/open?id=1WSmPaUV909uWF3ycL0MLWPWM6MdSjaJ0

Baidu Disk: https://pan.baidu.com/s/1f63o_QV4za6cdcigHSwhWw extraction code：jp53

Here include the saliency prediction maps for 37 major deep salient object detection (SOD) methods, a constructed dataset with annotations for attribute analysis, and codes for evaluation (see our paper for details).

1. Saliency prediction maps DUT.rar (DUT-OMRON dataset) DUTSTE.rar (test set of DUTS dataset) ECSSD.rar (ECSSD dataset) HKU-IS.rar (HKU-IS dataset) PASCAL-S.rar (PASCAL-S dataset) SOD.rar (SOD dataset)

2. Dataset and annotations for attribute analysis The hybrid dataset consists of 1,800 images randomly selected from 6 datasets, namely SOD, ECSSD, DUT-OMRON, PASCAL-S, HKU-IS and the test set of DUTS (300 for each). We carefully exclude images in ECSSD that also appear in SOD.

The annotations listed in ATTR_anno.xlsx cover 16 attributes from the perspectives of salient object categories, challenges and scene categories.

3. Codes for evaluation Matlab codes for calculating F-max, S-measure and MAE.

===========================================================================

Citation:

@article{wang2019sodsurvey,
	title={Salient Object Detection in the Deep Learning Era: An In-Depth Survey},
	author={Wang, Wenguan and Lai, Qiuxia and Fu, Huazhu and Shen, Jianbing and Ling, Haibin},
	journal={arXiv preprint arXiv:1904.09146},
	year={2019},
}

If you find our dataset is useful, please cite above paper.

===========================================================================

Contact Information

Email:

[email protected]

[email protected]