Skip to content

Latest commit

 

History

History
50 lines (37 loc) · 1.65 KB

README.md

File metadata and controls

50 lines (37 loc) · 1.65 KB
Raw

MFFN: Multi-view Feature Fusion Network for Camouflaged Object Detection

This is a PyTorch implementation of the MFFN paper:

@InProceedings{Zheng_2023_WACV, 
	author = {Dehua Zheng and Xiaochen Zheng and Laurence Yang and Yuan Gao and Chenlu Zhu and Yiheng Ruan}, 
	title = {MFFN: Multi-view Feature Fusion Network for Camouflaged Object Detection}, 
	booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)}, 
	year = {2023}, 
}

Installation

conda create -n MFFN python=3.8
conda activate MFFN
pip install torch==1.8.1 torchvision
git clone https://github.com/dwardzheng/MFFN_COD.git
cd MFFN_COD
pip install -r requirements.txt

Training

python main.py --model-name=MFFN --config=configs/MFFN/MFFN_R50.py --datasets-info ./configs/_base_/dataset/dataset_configs.json --info demo

Evaluation

./test.sh 0

Visualization

Visualization of camouflaged animal detection. The state-of-the-art and classic single-view COD model SINet is confused by the background sharing highly similarities with target objects and missed a lot of boundary and region shape information (indicated by orange arrows). Our multi-view scheme will eliminate these distractors and perform more efficiently and effectively.

License

This project is under the CC-BY-NC 4.0 license. See LICENSE for details.