Table 4. Performance benchmarking with... | Emerald Publishing

Table 4.

Performance benchmarking with weakly and fully supervised methods in the CryoNuSeg data set

Method	AJI	Dice	PQ
Weakly supervised
BoNuS (Lin et al., 2024)	0.431	0.693	0.399
Partial points (Qu et al., 2019)	0.410	0.682	0.357
DAWN (Zhang et al., 2024)	0.508	0.804	0.476
Pseudoedgenet (Yoo et al., 2019)	0.321	0.620	0.306
DoNuSeg (Wang et al., 2024)	0.441	0.672	0.306
Fully supervised
U-Net (Ronneberger et al., 2015)	0.469	0.697	0.403
HoVer-Net (Graham et al., 2019)	0.526	0.804	0.495
Swin-unet (Cao et al., 2022)	0.524	0.849	0.498
CDNet (He et al., 2021)	0.539	0.776	0.499
LG-NuSegHop (baseline)	0.545	0.703	0.419
LG-NuSegHop (dom. Adapted)	0.567	0.723	0.479

Method	AJI	Dice	PQ
Weakly supervised
BoNuS (Lin et al., 2024)	0.431	0.693	0.399
Partial points (Qu et al., 2019)	0.410	0.682	0.357
DAWN (Zhang et al., 2024)	0.508	0.804	0.476
Pseudoedgenet (Yoo et al., 2019)	0.321	0.620	0.306
DoNuSeg (Wang et al., 2024)	0.441	0.672	0.306
Fully supervised
U-Net (Ronneberger et al., 2015)	0.469	0.697	0.403
HoVer-Net (Graham et al., 2019)	0.526	0.804	0.495
Swin-unet (Cao et al., 2022)	0.524	0.849	0.498
CDNet (He et al., 2021)	0.539	0.776	0.499
LG-NuSegHop (baseline)	0.545	0.703	0.419
LG-NuSegHop (dom. Adapted)	0.567	0.723	0.479