https://github.com/wenguanwang/SODsurvey
Salient Object Detection in the Deep Learning Era: An In-Depth Survey
https://github.com/wenguanwang/SODsurvey
saliency saliency-maps saliency-prediction salient-object-detection sod survey visual-attention
Last synced: 6 months ago
JSON representation
Salient Object Detection in the Deep Learning Era: An In-Depth Survey
- Host: GitHub
- URL: https://github.com/wenguanwang/SODsurvey
- Owner: wenguanwang
- Created: 2019-04-08T08:23:18.000Z (about 7 years ago)
- Default Branch: master
- Last Pushed: 2021-01-09T11:56:25.000Z (over 5 years ago)
- Last Synced: 2024-06-29T04:36:48.554Z (almost 2 years ago)
- Topics: saliency, saliency-maps, saliency-prediction, salient-object-detection, sod, survey, visual-attention
- Homepage:
- Size: 14.8 MB
- Stars: 382
- Watchers: 10
- Forks: 69
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- awesome-segmentation-saliency-dataset - https://github.com/wenguanwang/SODsurvey
README
# [Salient object detection in the deep learning era: An in-depth survey, PAMI2021](https://www.researchgate.net/publication/332553805_Salient_Object_Detection_in_the_Deep_Learning_Era_An_In-Depth_Survey)
===========================================================================
[Wenguan Wang](https://sites.google.com/view/wenguanwang), Qiuxia Lai, Huazhu Fu, Jianbing Shen, Haibin Ling, Ruigang Yang
===========================================================================
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 46 major deep salient object detection (SOD) methods, a constructed dataset with annotations for attribute analysis, and codes for evaluation (see our paper for details).
===========================================================================
## :fire::fire::fire:Update
2020/1: Results of eight ICCV'19 methods are added.
2019/9: Results of eight CVPR'19 methods are added.
===========================================================================
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 and Yang, Ruigang},
journal={TPAMI},
year={2021},
}
If you find our dataset is useful, please cite above paper.
===========================================================================
Contact Information
Email:
wenguanwang.ai@gmail.com
qxlai@cse.cuhk.edu.hk