A novel deep learning-based model for automated tooth detection and numbering in mixed and permanent dentition in occlusal photographs

Table 2 Detailed performance metrics for all teeth

Tooth number	True positive (n)	False positive (n)	False negative (n)	Total teeth (n)	Sensitivity (%)	Precision (%)	F1score (%)
Upper 1	460	4	-	464	100	99.14	99.57
Upper 2	398	12	1	411	99.75	97.07	98.39
Upper 3	233	8	1	242	99.57	96.68	98.11
Upper 4	288	7	-	295	100	97.63	98.80
Upper 5	249	2	-	251	100	99.20	99.60
Upper 6	443	20	-	463	100	95.68	97.79
Upper 7	188	5	1	194	99.47	97.41	98.43
Upper 8	25	1	-	26	100	96.15	98.04
Lower 1	492	4	-	496	100	99.19	99.60
Lower 2	476	8	2	486	99.58	98.35	98.96
Lower 3	339	15	1	355	99.71	95.76	97.69
Lower 4	301	13	-	314	100	95.86	97.89
Lower 5	229	10	-	239	100	95.82	97.86
Lower 6	439	39	-	478	100	91.84	95.75
Lower 7	203	19	-	222	100	91.44	95.53
Lower 8	22	3	-	25	100	88	93.62
Upper A	1	2	-	3	100	33.33	50
Upper B	16	6	-	22	100	72.73	84.21
Upper C	162	21	-	183	100	88.52	93.91
Upper D	130	8	-	138	100	94.2	97.01
Upper E	193	11	-	204	100	94.61	97.23
Lower C	96	9	-	105	100	91.43	95.52
Lower D	131	10	-	141	100	92.91	96.32
Lower E	206	19	-	225	100	91.56	95.59

ISSN: 1472-6831