An automatic segmentation and classification framework for anti-nuclear antibody images

Table 4 Classification accuracies (%) of different cell patterns

Exp.	Iter.	All	Diffused	Peripheral	Nucleolar	Coarse-speckled	Fine-speckled	Discr.-speckled
1	-	99.69	98.33	100	98.94	99.95	100	100
2	1	97.65	93.98	99.24	82.98	98.88	85.71	99.33
	2	96.45	88.96	98.48	95.74	98.57	71.43	97.65
	3	97.23	93.65	97.72	87.23	98.47	78.57	99.66
	4	96.45	92.31	98.10	82.98	98.36	78.57	96.64
	5	97.13	94.98	96.96	89.36	98.47	71.43	98.66
	6	96.81	93.98	99.24	76.60	98.26	85.71	96.98
	7	96.81	94.31	97.34	87.23	98.67	82.14	95.64
	8	97.33	93.65	98.48	91.49	98.98	78.57	97.32
	9	96.50	90.30	98.48	93.62	98.47	85.71	95.97
	10	96.50	90.64	98.48	85.10	98.36	82.14	97.65
	Avg	96.87	92.68	98.25	87.23	98.55	80.00	97.55

Classified accuracies of 6 patterns of cells and overall cells achieved for CV and ITT experiments.

ISSN: 1475-925X