The more qualitative results of HitNet on four benchmarks (CAMO, CHAMELEON, NC4K, COD10K) have already been stored in Google Drive.
The pretrained model is stored in Google Drive. After downloading, please put it in the pretrained_pvt fold.
Our well-trained model is stored in Google Drive, which should be moved into the fold 'model_pth'. Google driver is out of service, please refer to Huggingface.If you have any quesetion, please let me know!
run sh train.sh for training, test.sh for testing.
@article{hu2022high,
title={High-resolution Iterative Feedback Network for Camouflaged Object Detection},
author={Hu, Xiaobin and Wang,Shuo and Qin, Xuebin and Dai, Hang and Ren, Wenqi and Tai, Ying and Wang, Chengjie and Shao, Ling},
journal={arXiv preprint arXiv:2203.11624},
year={2022}
}
@inproceedings{hu2023high,
title={High-resolution iterative feedback network for camouflaged object detection},
author={Hu, Xiaobin and Wang, Shuo and Qin, Xuebin and Dai, Hang and Ren, Wenqi and Luo, Donghao and Tai, Ying and Shao, Ling},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
volume={37},
number={1},
pages={881--889},
year={2023}
}