The training and testing experiments are conducted using PyTorch with 8 Tesla V100 GPUs of 36 GB Memory.
Note that FSPNet is only tested on Ubuntu OS with the following environments.
conda create -n FSPNet python=3.8
.pip install -r requirements.txt
train.sh
or slurm_train.sh
as needed to train.Our well-trained model is stored in Google Drive and Baidu Drive (otz5). After downloading, please change the file path in the corresponding code.
main.m
and just run it to generate the evaluation results.slurm_eval.py
in the run_slurm
folder for evaluation.The prediction results of our FSPNet are stored on Google Drive and Baidu Drive (ryzg) please check.
@inproceedings{Huang2023Feature,
title={Feature Shrinkage Pyramid for Camouflaged Object Detection with Transformers},
author={Huang, Zhou and Dai, Hang and Xiang, Tian-Zhu and Wang, Shuo and Chen, Huai-Xin and Qin, Jie and Xiong, Huan},
booktitle={IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2023}
}
Thanks to Deng-Ping Fan, Ge-Peng Ji, et al. for a series of efforts in the field of COD.