EGNet
EGNet:Edge Guidance Network for Salient Object Detection (ICCV 2019)
We use the sal2edge.m to generate the edge label for training.
For training:
-
Clone this code by
git clone https://github.com/JXingZhao/EGNet.git --recursive
, assume your source code directory is$EGNet
; -
Download training data (fsex) (google drive);
-
Download initial model (8ir7) (google_drive);
-
Change the image path and intial model path in run.py and dataset.py;
-
Start to train with
python3 run.py --mode train
.
For testing:
-
Download pretrained model (2cf5) (google drive);
-
Change the test image path in dataset.py
-
Generate saliency maps for SOD dataset by
python3 run.py --mode test --sal_mode s
, PASCALS bypython3 run.py --mode test --sal_mode p
and so on; -
Testing code we use is the public open source code. (https://github.com/Andrew-Qibin/SalMetric)
Pretrained models, datasets and results:
| Page | | Training Set (fsex) (google drive) | | Pretrained models (2cf5) | | Saliency maps (54gi) (google drive VGG) (google drive resnet) |
If you think this work is helpful, please cite
@inproceedings{zhao2019EGNet,
title={EGNet:Edge Guidance Network for Salient Object Detection},
author={Zhao, Jia-Xing and Liu, Jiang-Jiang and Fan, Deng-Ping and Cao, Yang and Yang, Jufeng and Cheng, Ming-Ming},
booktitle={The IEEE International Conference on Computer Vision (ICCV)},
month={Oct},
year={2019},
}
Other related work
Contrast Prior and Fluid Pyramid Integration for RGBD Salient Object Detection. (CVPR2019) page