Survey statistics of automated segmentations applied to optical imaging of mammalian cells

BMC Bioinformatics

Table 9 A summary of segmentation evaluation metrics

Measures based on	Metric name	Cellular measurement	Reference
Number of Mis-segmented voxels	Jaccard	Synthetic	[65]
	Dice	Cell	[120, 129, 154]
		Synthetic	[154]
		Other	[66]
	F-Measure	Synthetic	[155]
	Adjusted Rand Index	Cell	[122]
	Custom measure	Nucleus	[61]
	Custom measure	Cell	[67]
	Misclassification error	Nucleus	[156]
	Misclassification error	Other	[156]
	Accuracy (ACC)	Cell	[157, 158]
Position of mis-segmented voxels	Average distance	Cell	[56]
		Synthetic	[117]
		Other	[116]
	Root square mean of deviation	Synthetic	[159]
	Histogram of distances	Nucleus	[138]
Number of objects	Object count	Nucleus	[55, 56, 123, 160–162]
	Object count	Cell	[81, 119, 163]
	Precision/Recall	Nucleus	[54, 84]
	Precision/Recall	Cell	[44, 69, 84, 127]
	F-measure	Nucleus	[84]
	F-measure	Cell	[69, 84]
	Bias index	Cell	[69]
	Sensitivity	Nucleus	[138, 164]
	Custom measure	Cell	[67]
	Cell detection rate	Cell	[165]
Feature values of segmented objects	Velocity histogram	Cell	[166]
	Object position	Nucleus	[167]
		Cell	[151, 163, 166]
		Synthetic	[168]
	Pearson’s correlation slope and intercept for velocity measurements	Cell	[166]
	Voxel intensity based	Synthetic	[159]
	Voxel intensity based	Other	[73]
	Object area and shape based	Cell	[151]
	Object area and shape based	Other	[73]
	Structural index	Cell	[151]

ISSN: 1471-2105