Feature Quantification and Abnormal Detection on Cervical Squamous Epithelial Cells

Table 2

Abnormal cell detection based on AP algorithm in RGB color space.

(a)

Clustering	Distance of samples
	Criterion: the value in RGB color space
Categories	I	II	III	IV	V
Number of clustering centers	34	14	19	33	7
Samples in the class	1, 2, 10, 24, 25, 23	9, 13, 17, 28, 29, 31, 32	3, 4, 6, 8, 11, 12, 16, 18, 20, 27, 30	5, 21, 22, 26	15
Scope of feature values
Feature thresholds

(b)

Clustering	Distance of samples
	Criterion: the value in RGB color space
Categories	I	II	III	IV
Number of clustering centers	31	2	33	3
Samples in the class	18, 23, 24, 25, 27, 32	1, 10, 21, 22, 28, 34	4, 5, 9, 13, 14, 20, 29	6, 7, 8, 11, 12, 15, 16, 17, 19, 26, 30
Scope of feature values
Feature thresholds

(c)

Clustering	Distance of samples
	Criterion: the value in RGB color space
Categories	I	II	III	IV	V
Number of clustering centers	18	25	14	20	19
Samples in the class	23, 24, 27	2, 10, 21, 22, 28, 31, 32, 34	1, 5	4, 9, 13, 16, 26, 29, 33	3, 6, 7, 8, 11, 12, 15, 17, 30
Scope of feature values
Feature thresholds