Skip to content

rmcong/CIRNet_TIP2022

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

31 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CIRNet_TIP2022

Runmin Cong, Qinwei Lin, Chen Zhang, Chongyi Li, Xiaochun Cao, Qingming Huang, and Yao Zhao, CIR-Net: Cross-modality interaction and refinement for RGB-D salient object detection, IEEE Transactions on Image Processing, vol. 31, pp. 6800-6815, 2022.

Results of CIR-Net

  • Results:
  • We provide the resutls of our CIR-Net on six popular RGB-D SOD benchmark datasets, including STEREO797, NLPR, NJUD, DUT, LFSD and SIP.
  • The results can be download from: Baidu Cloud (password:1234)

Pytorch Code of CIR-Net:

  • Pytorch implementation of CIR-Net
  • Pretrained model:
    • We provide our testing code. If you test our model, please download the pretrained model, unzip it, and put the checkpoint CIRNet.pth to CIRNet_cpts/ folder
    • Pretrained model using ResNet50 backbone:Baidu Cloud (password:1234)
    • Pretrained model using VGG16 backbone: Baidu Cloud (password:1234)

Requirements

  • Python 3.7
  • torch=1.10.1
  • torchvision=0.11.2
  • opencv-python
  • Pillow

Data Preprocessing

  • Please download and put the train data to data folder.
  • train data can be download from: Baidu Cloud (password:1234)
  • test data can be download from: Baidu Cloud (password:1234)

Test

python3 CIRNet_test.py --backbone R50 --test_model CIRNet_R50.pth

Train

python3 CIRNet_train.py --backbone R50
  • You can find the results in the test_maps folder

If you use our CIR-Net, please cite our paper:

 @article{crm/tip22/CIRNet,
   title={{CIR-Net}: Cross-modality interaction and refinement for {RGB-D} salient object detection},
   author={Cong, Runmin and Lin, Qinwei and Zhang, Chen and Li, Chongyi and Cao, Xiaochun and Huang, Qingming and Zhao, Yao },
   journal={IEEE Trans. Image Process. },
   volume={31},
   pages={6800-6815},
   year={2022},
  }

Contact Us

If you have any questions, please contact Runmin Cong (rmcong@bjtu.edu.cn) or Qinwei Lin (lqw22@mails.tsinghua.edu.cn).

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages