An Evaluation of Cellular Neural Networks for the Automatic Identification of Cephalometric Landmarks on Digital Images

Table 2

Mean difference between the average distance from the best estimate of landmark position of automated and manual landmarking, standard deviation (SD) (expressed in mm), and results of 1-way ANOVA test. NS indicates no statistically significant difference and the number of asterisks (*) the level of significance.

Landmark coordinates		Mean difference	Standard deviation (SD)	𝐹-value	𝑃-value (ANOVA)	S
Nasion	𝑋	0.217	±0.441	5.076	.027	*
	𝑌	0.483	±0.442	16.229	.000	**
A point	𝑋	0.596	±0.312	46.633	.000	***
	𝑌	0.101	±0.282	0.484	.489	NS
Basion	𝑋	0.181	±0.197	2.160	.146	NS
	𝑌	0.225	±0.112	2.538	.115	NS
Porion	𝑋	0.003	±0.145	0.000	.986	NS
	𝑌	0.538	±0.257	24.051	.000	***
Pterygoid point	𝑋	0.157	±0.296	1.786	.185	NS
	𝑌	0.022	±0.033	0.0298	.863	NS
B point	𝑋	0.161	±0.130	6.846	.011	*
	𝑌	0.285	±0.402	2.762	.101	NS
Pogonion	𝑋	0.038	±0.338	0.238	.627	NS
	𝑌	0.166	±0.033	1.385	.243	NS
Protuberance Menti	𝑋	0.038	±0.226	0.314	.577	NS
	𝑌	0.245	±0.139	2.457	.121	NS
Upper incisor edge	𝑋	0.172	±0.030	4.712	.033	*
	𝑌	0.100	±0.371	0.283	.596	NS
Lower incisor edge	𝑋	0.226	±0.226	4.582	.035	*
	𝑌	0.283	±0.089	3.796	.055	NS

*Significance level .05. ^**Significance level .01. ^***Significance level .001.