Computer-aided design and 3-dimensional artificial/convolutional neural network for digital partial dental crown synthesis and validation

Abstract

The current multiphase, invitro study developed and validated a 3-dimensional convolutional neural network (3D-CNN) to generate partial dental crowns (PDC) for use in restorative dentistry. The effectiveness of desktop laser and intraoral scanners in generating data for the purpose of 3D-CNN was first evaluated (phase 1). There were no significant differences in surface area [t-stat(df) = − 0.01 (10), mean difference = − 0.058, P > 0.99] and volume [t-stat(df) = 0.357(10)]. However, the intraoral scans were chosen for phase 2 as they produced a greater level of volumetric details (343.83 ± 43.52 mm3) compared to desktop laser scanning (322.70 ± 40.15 mm3). In phase 2, 120 tooth preparations were digitally synthesized from intraoral scans, and two clinicians designed the respective PDCs using computer-aided design (CAD) workflows on a personal computer setup. Statistical comparison by 3-factor ANOVA demonstrated significant differences in surface area (P < 0.001), volume (P < 0.001), and spatial overlap (P < 0.001), and therefore only the most accurate PDCs (n = 30) were picked to train the neural network (Phase 3). The current 3D-CNN produced a validation accuracy of 60%, validation loss of 0.68–0.87, sensitivity of 1.00, precision of 0.50–0.83, and serves as a proof-of-concept that 3D-CNN can predict and generate PDC prostheses in CAD for restorative dentistry.

Revolutionizing CAD/CAM-based restorative dental processes and materials with artificial intelligence: a concise narrative review

“…In 2023, Chau et al (2023) and Ding et al (2023a) demonstrated that the 3D GAN and 3D-DCGAN networks are beneficial in automatically designing biomimetic dental restorations by learning the occlusal anatomic features of similar remaining teeth. Farook et al (2023) created a new 3D dental prosthetic dataset for 3D convolutional neural network (3D-CNN) training purposes, as virtual data were successfully used in both open-source and commercial CAD workflows to create tooth preparations for machine learning (ML) in restorative dentistry. 3D deep learning (DL) was able to accurately produce inlays and onlays for suitable tooth preparations ( Farook et al, 2023 ).…”

“… Farook et al (2023) created a new 3D dental prosthetic dataset for 3D convolutional neural network (3D-CNN) training purposes, as virtual data were successfully used in both open-source and commercial CAD workflows to create tooth preparations for machine learning (ML) in restorative dentistry. 3D deep learning (DL) was able to accurately produce inlays and onlays for suitable tooth preparations ( Farook et al, 2023 ).…”

The Potential of Artificial Intelligence in Prosthodontics: A Comprehensive Review

“…Results of the Quadas-2 assessment of included studies are presented in Table 6. A total of 18.3% of studies (7/38) were identified as low-risk of bias studies [8,25,29,33,43,45,46], whereas 52.6% (20/38) and 28.9% (11/38) of included studies were identified as studies with high [12-15,17-19,20-22,26,28,30,34-38, 42,44] and unclear [4,5,16,23,24,27,31,32,39,40,46] risk of bias, respectively. Inter-rater and intra-rater reliability of the Cohen kappa were 0.8831 and 0.9232, respectively, which reflect near-perfect reliability.…”

Teeth Microcracks Research: Towards Multi-Modal Imaging

“…Initially, studies focused on traditional convolutional neural networks (CNNs) such as U-Net for caries detection and localization [74,75]. Subsequent progress led to the development of more sophisticated architectures, including a 3D-CNN for generating partial dental crowns, demonstrating improved validation accuracy and sensitivity [76]. The evolution continued with the introduction of self-training-based methods, leveraging unlabeled data for student model training, offering computational and performance gains over traditional supervised learning [77].…”