Camouflaged Object Detection for Activated Sludge Microorganisms Based on Context-Aware

Table 1. Settings of image number for different datasets
View table
Table 1. Settings of image number for different datasets
Dataset Total Training set Test set
CAMO 1250 1000 250
COD10K 5066 3040 2026
CHAMELEON 76 0 76
NC4K 4121 0 4121
CASM 1888 1321 567

Table 2. Comparison of performance indicators for different models on four datasets

View table

Table 2. Comparison of performance indicators for different models on four datasets

Model	CAMO-Test				COD10K-Test
Model	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓
SINet	0.745	0.712	0.825	0.092	0.776	0.667	0.867	0.043
SINet-V2	0.811	0.733	0.871	0.074	0.812	0.678	0.866	0.037
SegMaR	0.815	0.795	0.881	0.071	0.833	0.739	0.893	0.034
BGNet	0.813	0.786	0.876	0.073	0.831	0.739	0.902	0.033
HitNet	0.825	0.768	0.895	0.062	0.819	0.711	0.899	0.033
FSEL	0.821	0.799	0.893	0.067	0.830	0.743	0.898	0.031
Proposed	0.863	0.812	0.922	0.050	0.851	0.750	0.906	0.027
Model	CHAMELEON				NC4K
Model	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓
SINet	0.872	0.823	0.938	0.034	0.808	0.768	0.883	0.058
SINet-V2	0.885	0.815	0.928	0.033	0.846	0.768	0.902	0.048
SegMaR	0.906	0.866	0.952	0.025	0.841	0.821	0.905	0.046
BGNet	0.901	0.846	0.938	0.027	0.851	0.813	0.911	0.044
HitNet	0.883	0.836	0.935	0.030	0.851	0.796	0.914	0.043
FSEL	—	—	—	—	0.854	0.821	0.915	0.042
Proposed	0.902	0.848	0.946	0.025	0.879	0.825	0.929	0.035

Table 3. Comparison of performance indicators for different models on the CASM dataset

View table

Table 3. Comparison of performance indicators for different models on the CASM dataset

Model	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓
SINet	0.822	0.540	0.840	0.019
SINet-V2	0.925	0.871	0.960	0.007
SegMaR	0.930	0.877	0.963	0.008
BGNet	0.934	0.868	0.964	0.009
HitNet	0.893	0.814	0.939	0.011
FSEL	0.895	0.801	0.915	0.011
Proposed	0.935	0.887	0.970	0.006

Table 4. Ablation experimental results on four public datasets

View table

Table 4. Ablation experimental results on four public datasets

Experiment	MCCS	PVT v2	CAMO-Test				COD10K-Test
Experiment	MCCS	PVT v2	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓
1			0.811	0.733	0.1	0.074	0.812	0.678	0.866	0.037
2	✓		0.843	0.756	0.892	0.061	0.822	0.703	0.873	0.035
3		✓	0.861	0.807	0.920	0.052	0.846	0.738	0.894	0.028
4 （proposed）	✓	✓	0.863	0.812	0.922	0.050	0.851	0.750	0.906	0.027
Experiment	MCCS	PVT v2	CHAMELEON				NC4K
Experiment	MCCS	PVT v2	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓	$S_{α}$ ↑	$F_{β}^{w}$ ↑	$E_{ϕ}$ ↑	$M A E$ ↓
1			0.885	0.815	0.928	0.033	0.846	0.768	0.902	0.048
2	✓		0.890	0.827	0.931	0.030	0.859	0.791	0.910	0.041
3		✓	0.898	0.840	0.939	0.025	0.877	0.819	0.924	0.037
4 （proposed）	✓	✓	0.902	0.848	0.946	0.025	0.879	0.825	0.929	0.035

Table 5. Ablation experimental results on the CASM dataset
View table
Table 5. Ablation experimental results on the CASM dataset
Experiment MCCS PVT v2 $S_{α}$ ↑ $F_{β}^{w}$ ↑ $E_{ϕ}$ ↑ $M$ ↓
1 0.925 0.871 0.960 0.007
2 ✓ 0.932 0.883 0.966 0.007
3 ✓ 0.927 0.875 0.963 0.007
4（proposed） ✓ ✓ 0.935 0.887 0.970 0.006

Tools

Get Citation

Copy Citation Text

Yifan Fang, Lijie Zhao, Mingxi Jin, Mingzhong Huang, Zhenpeng Gao. Camouflaged Object Detection for Activated Sludge Microorganisms Based on Context-Aware[J]. Laser & Optoelectronics Progress, 2025, 62(16): 1618001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites