Trichoscopy of Alopecia Areata: Hair Loss Feature Extraction and Computation Using Grid Line Selection and Eigenvalue

Table 1

Evaluation of Input data in Figure 8. Bold data are assumed as bad experiment results and italic data as good experiment results.

	Input (a)	Input (b)	Input (c)	Input (d)	Input (e)	Input (f)	Dataset
Hair truth	18	20	23	23	17	25	19.59 (avg)
Hair prediction	15.42	22.49	22.73	12.97	15.21	23.12	19.04 (avg)
Hair difference	2.59	2.49	0.27	10.03	1.79	1.88	3.13
Follicle truth	9	8	7	7	8	7	8.15 (avg)
Follicle prediction	9	12	8	7	10	6	8.49 (avg)
Follicle difference	0	4	1	0	2	1	1.44
Thickness truth	53.8168	67.0311	72.8527	76.7309	66.6854	60.6011	—
Thickness prediction	61.0024	65.3314	67.3158	87.0331	58.4213	60.4601	61.98
Thickness difference	7.1856	1.6997	5.5369	10.3022	8.2641	0.141	—
Total accuracy	90.75	78.34	92.31	80.99	84.03	92.65	96.51 (avg)
							83.20 (diff)