Table 4.

The existing DL models on tooth segmentation and their functions and performance.

Authors	DL Models	Year	Training Dataset	Validation/Test Dataset	Functions	Best Performance of DL	Time- Consuming
Jin et al. [48]	Unknown	2022	216	223	Tooth identification and segmentation	Tooth identification Precision: 0.9681 ±  0.0167 Recall: 0.9013 ±  0.0530 F1 score: 0.9335 ±  0.0254 Tooth segment Precision: 0.9595 ±  0.0200 Recall: 0.9371 ±  0.0208 DSC: 0.9479 ±  0.0134 HD: 1.66 ±  0.72 mm	No
He et al. [49]	cGAN	2020	15,750 teeth	4200 teeth	Tooth identification and segmentation	IoU Incisor: 0.89 ± 0.087 Lateral incisor: 0.92 ± 0.068 Canine: 0.90 ± 0.053 First premolar: 0.91 ± 0.032 Second premolar: 0.93 ± 0.026 First molar: 0.92 ± 0.112 Second molar: 0.90 ± 0.035	No
Jacobs et al. [50]	CNN	2021	2095 slice	328 for validation 501 for optimization	Tooth segmentation	R-AI IoU: 0.881 ± 0.036 DSC: 0.937 ± 0.02 F-AI IoU: 0.887 ± 0.032 DSC: 0.940 ± 0.018	R-AI 72 ± 33.02 s F-AI 30 ± 8.64 s
Jacobs et al. [51]	3D U-Net	2021	140	35 for validation 11 for test	Tooth identification and segmentation	Precision: 0.98 ± 0.02 IoU: 0.82 ± 0.05 Recall: 0.83 ± 0.05 DSC: 0.90 ± 0.03 95HD: 0.56 ± 0.38 mm	7 ± 1.2 h for experts 13.7 ± 1.2 s for DL
Deng et al. [52]	CNN	2022	450	104	Tooth identification and segmentation	Accuracy: 0.913 AUC: 0.997	No
Jacobs et al. [53]	CNN	2022	140	35	Tooth identification and segmentation	Accuracy of teeth detection: 0.997 Accuracy of missing teeth detection: 0.99 IoU: 0.96 95HD: 0.33	1.5 s
Ozyurek et al. [55]	CNN	2020	2800	153	Periapical pathosis detection and their volumes calculation	Detection rate: 0.928	No
Li et al. [56]	U-Net	2020	61	12	Periapical lesion, tooth, bone, material segmentation	Accuracy: 0.93 Specificity: 0.88 DSC: 0.78	No
Schwendicke et al. [58]	Xception U-Net	2021	100	35	Detect the C-shaped root canal of the second molar	DSC: 0.768 ± 0.0349 Sensitivity: 0.786 ± 0.0378	No
Mahdian et al. [59]	U-Net	2022	90	10	Unobturated mesial buccal 2 (MB2) canals on endodontically obturated maxillary molars	Accuracy: 0.9 DSC: 0.768 Sensitivity: 0.8 Specificity: 1	No
Xie et al [60]	cGAN	2021	Improved group 40 Traditional group 40		Different tooth parts segmentation	Omit, Precision, TRP, FRP, and DSC	No
Yang et al. [61]	RPN, FRN, U-Net	2021	20		Tooth and pulp segmentation	Single root tooth DSC: 0.957 ± 0.005 ASD: 0.104 ± 0.019 mm RVD: 0.049 ± 0.017 Multiroot tooth DSC: 0.962 ± 0.002 ASD: 0.137 ± 0.019 mm RVD: 0.053 ± 0.010	No
Lin et al. [62]	U-Net, AGs, RNN	2020	1160	361	Root segmentation	IoU: 0.914 DSC: 0.955 Precision: 0.958 Recall: 0.953	No
Lin et al. [63]	ResNet50, VGG19, DenseNet169	2022	839	279	Vertical root fracture diagnosis	ResNet50 Accuracy: 0.978 Sensitivity: 0.970 Specificity: 0.985 VGG19 Accuracy: 0.949 Sensitivity: 0.927 Specificity: 0.970 DenseNet169 Accuracy: 0.963 Sensitivity: 0.941 Specificity: 0.985	No
Zhao et al. [64]	3D U-Net	2021	51	17	Root canal system detection	DSC: 0.952	350 ms