ZoomNeXt

ZoomNeXt: A Unified Collaborative Pyramid Network for Camouflaged Object Detection (preprint)

Note

The paper is still under review. Considering that it may need to be further revised according to the reviewers' comments, the code cannot be stabilized at the moment. So it will still take some time for the full code to be released.

@misc{ZoomNeXt,
      title={ZoomNeXt: A Unified Collaborative Pyramid Network for Camouflaged Object Detection}, 
      author={Youwei Pang and Xiaoqi Zhao and Tian-Zhu Xiang and Lihe Zhang and Huchuan Lu},
      year={2023},
      eprint={2310.20208},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Results

Datasets
CAMO-TE, CHAMELEON, COD10K-TE, NC4K	ResNet-50, EfficientNet-B4, PVTv2-B2, PVTv2-B3, PVTv2-B4, PVTv2-B5
CAD, MoCA-Mask-TE	PVTv2-B5

CAMO-TE

Backbone	$S_m$	$F^{\omega}_{\beta}$	MAE
ResNet-50	0.833	0.774	0.065
EfficientNet-B4	0.867	0.824	0.046
PVTv2-B2	0.874	0.839	0.047
PVTv2-B3	0.885	0.854	0.042
PVTv2-B4	0.888	0.859	0.040
PVTv2-B5	0.889	0.857	0.041

CHAMELEON

Backbone	$S_m$	$F^{\omega}_{\beta}$	MAE
ResNet-50	0.908	0.858	0.021
EfficientNet-B4	0.911	0.865	0.020
PVTv2-B2	0.922	0.884	0.017
PVTv2-B3	0.927	0.898	0.017
PVTv2-B4	0.925	0.897	0.016
PVTv2-B5	0.924	0.885	0.018

COD10K-TE

Backbone	$S_m$	$F^{\omega}_{\beta}$	MAE
ResNet-50	0.861	0.768	0.026
EfficientNet-B4	0.875	0.797	0.021
PVTv2-B2	0.887	0.818	0.019
PVTv2-B3	0.895	0.829	0.018
PVTv2-B4	0.898	0.838	0.017
PVTv2-B5	0.898	0.827	0.018

NC4K

Backbone	$S_m$	$F^{\omega}_{\beta}$	MAE
ResNet-50	0.874	0.816	0.037
EfficientNet-B4	0.884	0.837	0.032
PVTv2-B2	0.892	0.852	0.030
PVTv2-B3	0.900	0.861	0.028
PVTv2-B4	0.900	0.865	0.028
PVTv2-B5	0.903	0.863	0.028

CAD

Backbone	$S_m$	$F^{\omega}_{\beta}$	MAE	mDice	mIoU
PVTv2-B5	0.757	0.593	0.020	0.599	0.510

MoCA-Mask-TE

Backbone	$S_m$	$F^{\omega}_{\beta}$	MAE	mDice	mIoU
PVTv2-B5	0.734	0.476	0.010	0.497	0.422