Multiple Morphological Constraints-Based Complex Gland Segmentation in Colorectal Cancer Pathology Image Analysis

Table 1

Comparing results of different competition algorithms on the public GlaS dataset.

Method	F1 score		Object Dice		Object Hausdorff
	Test A	Test B	Test A	Test B	Test A	Test B
Proposed-N + L	0.901	0.851	0.893	0.842	44.125	94.528
Proposed-N	0.886	0.816	0.886	0.823	45.236	103.686
CUMedVision1	0.868	0.769	0.867	0.800	74.596	153.646
CUMedVision2	0.912	0.716	0.897	0.781	45.418	160.347
ExB1	0.891	0.703	0.882	0.786	57.413	145.575
ExB2	0.892	0.686	0.884	0.754	54.785	187.442
ExB3	0.896	0.719	0.886	0.765	57.350	159.873
Freiburg1	0.834	0.605	0.875	0.783	57.194	146.607
Freiburg2	0.870	0.695	0.876	0.786	57.093	148.463
CVML	0.652	0.541	0.644	0.654	155.433	176.244
LIB	0.777	0.306	0.781	0.617	112.706	190.447
Vision4GlaS	0.635	0.527	0.737	0.610	107.491	210.105