Development of an artificial intelligence-based algorithm to classify images acquired with an intraoral scanner of individual molar teeth into three categories

Background Forensic dentistry identifies deceased individuals by comparing postmortem dental charts, oral-cavity pictures and dental X-ray images with antemortem records. However, conventional forensic dentistry methods are time-consuming and thus unable to rapidly identify large numbers of victims following a large-scale disaster. Objective Our goal is to automate the dental filing process by using intraoral scanner images. In this study, we generated and evaluated an artificial intelligence-based algorithm that classified images of individual molar teeth into three categories: (1) full metallic crown (FMC); (2) partial metallic restoration (In); or (3) sound tooth, carious tooth or non-metallic restoration (CNMR). Methods A pre-trained model was created using oral-cavity pictures from patients. Then, the algorithm was generated through transfer learning and training with images acquired from cadavers by intraoral scanning. Cross-validation was performed to reduce bias. The ability of the model to classify molar teeth into the three categories (FMC, In or CNMR) was evaluated using four criteria: precision, recall, F-measure and overall accuracy. Results The average value (variance) was 0.952 (0.000140) for recall, 0.957 (0.0000614) for precision, 0.952 (0.000145) for F-measure, and 0.952 (0.000142) for overall accuracy when the algorithm was used to classify images of molar teeth acquired from cadavers by intraoral scanning. Conclusion We have created an artificial intelligence-based algorithm that analyzes images acquired with an intraoral scanner and classifies molar teeth into one of three types (FMC, In or CNMR) based on the presence/absence of metallic restorations. Furthermore, the accuracy of the algorithm reached about 95%. This algorithm was constructed as a first step toward the development of an automated system that generates dental charts from images acquired by an intraoral scanner. The availability of such a system would greatly increase the efficiency of personal identification in the event of a major disaster.

Accuracy of full-arch digitalization for partially edentulous jaws — a laboratory study on basis of coordinate-based data analysis

Objectives To compare the accuracy (trueness and precision) of direct digitization of four different dental gap situation with two IOS (intraoral scanner). Materials and methods Four partially edentulous polyurethane mandible models were used: (1) A (46, 45, 44 missing), (2) B (45, 44, 34, 35 missing), (3) C (42, 41, 31, 32 missing), and (4) D (full dentition). On each model, the same reference object was fixed between the second molars of both quadrants. A dataset (REF) of the reference object was generated by a coordinate measuring machine. Each model situation was scanned by (1) OMN (Cerec AC Omnicam) and (2) PRI (Cerec Primescan AC) (n = 30). Datasets of all 8 test groups (N = 240) were analyzed using inspection software to determine the linear aberrations in the X-, Y-, Z-axes and angular deviations. Mann–Whitney U and two-sample Kolmogorov–Smirnov tests were used to detect differences for trueness and precision. Results PRI revealed higher trueness and precision in most of the measured parameters ($${\overrightarrow{V}}_{E}$$ V → E 120.95 to 175.01 μm, $$\overrightarrow{V}_{E}$$ V → E (x) − 58.50 to − 9.40 μm, $$\overrightarrow{V}_{E}$$ V → E (z) − 70.35 to 63.50 μm), while OMN showed higher trueness for $$\overrightarrow{V}_{E}$$ V → E (y) regardless of model situation (− 104.90 to 34.55 μm). Model D revealed the highest trueness and precision in most of the measured parameters regardless of IOS ($$\overrightarrow{V}_{E}$$ V → E 120.95 to 195.74 μm, $$\overrightarrow{V}_{E}$$ V → E (x) − 9.40 to 66.75 μm,$$\overrightarrow{V}_{E}$$ V → E (y) − 14.55 to 51.50 μm, $$\overrightarrow{V}_{E}$$ V → E (z) 63.50 to 120.75 μm). Conclusions PRI demonstrated higher accuracy in the X- and Z-axes, while OMN depicted higher trueness in the Y-axis. For PRI, Model A revealed the highest distortion, while for OMN, Model B produced the largest aberrations in most parameters. Clinical relevance Current results suggest that both investigated IOS are sufficiently accurate for the manufacturing of tooth-borne restorations and orthodontic appliances. However, both hardware specifications of IOS and the presence of edentulous gaps in the dental model have an influence on the accuracy of the virtual model dataset.

Accuracy of photogrammetry, intraoral scanning, and conventional impression techniques for complete-arch implant rehabilitation: an in vitro comparative study

Background To compare the accuracy of photogrammetry, intraoral scanning and conventional impression techniques for complete-arch implant rehabilitation. Methods A master cast containing 6 implant abutment replicas was fabricated. Group PG: digital impressions were taken 10 times using a photogrammetry system; Group IOS: intraoral scanning was performed to fabricate 10 digital impressions; Group CNV: splinted open-tray impression technique was used to fabricate 10 definitive casts. The master cast and conventional definitive casts were digitized with a laboratory reference scanner. For all STL files obtained, scan bodies were converted to implant abutment replicas using a digital library. The accuracy of a digitizer was defined by 2 main parameters, trueness and precision. "Trueness" was used to describe the deviation between test files and reference file, and "precision" was used to describe the closeness between test files. Then, the trueness and precision of three impression techniques were evaluated and statistically compared (α = 0.05). Results The median trueness was 24.45, 43.45 and 28.70 μm for group PG, IOS and CNV; Group PG gave more accurate trueness than group IOS (P < 0.001) and group CNV (P = 0.033), group CNV showed more accurate trueness than group IOS (P = 0.033). The median precision was 2.00, 36.00 and 29.40 μm for group PG, IOS and CNV; Group PG gave more accurate precision than group IOS (P < 0.001) and group CNV (P < 0.001), group CNV showed more accurate precision than IOS (P = 0.002). Conclusions For complete-arch implant rehabilitation, the photogrammetry system showed the best accuracy of all the impression techniques evaluated, followed by the conventional impression technique, and the intraoral scanner provided the least accuracy.

The Accuracy and Reliability of Tooth Shade Selection Using Different Instrumental Techniques: An In Vitro Study

This study aimed to investigate and compare the reliability and accuracy of tooth shade selection in the model using 30 milled crowns via five methods: (1) digital single-lens reflex (DSLR) camera with twin flash (TF) and polarized filter (DSLR + TF), (2) DSLR camera with a ring flash (RF) and polarized filter (DSLR + RF), (3) smartphone camera with light corrector and polarized filter (SMART), (4) intraoral scanner (IOS), and (5) spectrophotometer (SPEC). These methods were compared with the control group or manufacturer’s shade. The CIE Lab values (L, a, and b values) were obtained from five of the methods to indicate the color of the tooth. Adobe Photoshop was used to generate CIE Lab values from the digital photographs. The reliability was calculated from the intraclass correlation based on two repetitions. The accuracy was calculated from; (a) ΔE calculated by the formula comparing each method to the control group, (b) study and control groups were analyzed by using the Kruskal–Wallis test, and (c) the relationship between study and control groups were calculated using Spearman’s correlation. The reliability of the intraclass correlation of L, a, and b values obtained from the five methods showed satisfactory correlations ranging from 0.732–0.996, 0.887–0.994, and 0.884–0.999, respectively. The ΔE from all groups had statistically significant differences when compared to the border of clinical acceptance (ΔE = 6.8). The ΔE from DSLR + TF, DSLR + RF, SMART, and SPEC were higher than clinical acceptance (ΔE > 6.8), whereas the ΔE from IOS was 5.96 and all of the L, a, and b values were not statistically significantly different from the manufacturer’s shade (p < 0.01). The ΔE of the DSLR + RF group showed the least accuracy (ΔE = 19.98), whereas the ∆E of DSLR + TF, SMART, and SPEC showed similar accuracy ∆E (ΔE = 10.90, 10.57, and 11.57, respectively). The DSLR camera combined with a ring flash system and polarized filter provided the least accuracy. The intraoral scanner provided the highest accuracy. However, tooth shade selection deserves the combination of various techniques and a professional learning curve to establish the most accurate outcome.

Accuracy of the Intra- and Extra-Oral Scanning Technique for Transferring the Intaglio Surface of a Pontic of Provisional Restorations to Definitive Restorations

When taking the final impression for a three-unit fixed partial denture (FPD), the intaglio surface of the pontic of provisional restoration cannot be transferred accurately to that of definitive restoration. The intra- and extra-oral scanning (IEOS) technique, a method for accurately reproducing the submucosal morphology of the superstructure of an implant, has been reported using an intraoral scanner. In the present study, we evaluated the difference between the conventional impression method using impression material and the IEOS technique in reproducing the morphology of the surface of the pontic of a definitive FPD. There was a significant difference in the trueness of the intaglio surface morphology of the pontic between the conventional method and the IEOS technique; however, no significant difference in precision was observed. As a result, the intaglio surface of the pontic of the three-unit FPD could be transferred to definitive restorations more accurately with the IEOS technique than with the conventional method. These results suggest that the IEOS technique can duplicate the intaglio surface of the pontic more reproducibly to the definitive restorations compared with the conventional method.

Virtual occlusal record: a literature review about the digital method

To produce prostheses through the digital flow, it is essential to transfer the correct patient’s interocclusal relationship to the digital software program, enabling the articulation of virtual models. Therefore, the aim of this study was to carry out a narrative literature review to describe and discuss aspects related to the virtual occlusal record realization, as well as its precision and accuracy in different clinical situations. Searches for scientific publications were performed in different databases and only articles in English related to the topic were selected. Different methods for the alignment of virtual models are described in the literature, the main one being the scanning of the patient in occlusion, usually in a position of maximum intercuspation. However, this technique may demonstrate disagreement with the patient's actual occlusal relationship due to several factors, and therefore studies were carried out to verify the precision and accuracy of these records. Most studies use plaster models and industrial scanner to capture the record, with few studies performed with intraoral scanner in patients. Despite the various scanner systems available and the different ways of evaluating them, in general, the studies show an adequate precision and accuracy of virtual occlusal records of dentate models. However, the absence of dental elements is related to the lower accuracy of these records, it being necessary to establish an appropriate method of scanning for these clinical situations.

Comparison of the Accuracy of Intraoral Scanners Based on the Type of Tooth Preparation for a Single Crown

The purpose of this study was to evaluate the accuracy of intraoral scanners in 10 abutments (five premolars and five molars) obtained in a dental clinic and to analyze the impacts of the volume and area of abutments on scanning accuracy. Abutment casts were scanned five times with a 3D contact scanner (DS10; Renishaw plc). The five scan files were lined up and then merged, and one high-resolution computer-aided design reference model (CRM) was obtained. To obtain a computer-aided design test model (CTM), three types of intraoral scanners (CS3600 (Carestream Dental), i500 (Medit), and EZIS PO (DDS)) and one type of laboratory scanner (E1; 3Shape) were employed. Using 3D analysis software (Geomagic control X; 3D Systems), the accuracy of the scanners was evaluated, including optimal overlap by optimal alignment. The conformity of the overlapped data was calculated by the root mean square (RMS) value, using the 3D compare function for evaluation. As for statistical analysis, testing was conducted, using one-way and two-way ANOVA and the Tukey HSD test (α = 0.05) for the comparison of the groups. To analyze the correlations of the volume and area of the abutments with accuracy, Pearson’s correlation analysis was conducted (α = 0.00625). Both premolar and molar abutments showed a lower RMS value on the laboratory scanner than on the intraoral scanners, and the RMS value was lower in premolars than in molars (p < 0.001). In the intraoral scanner group, CS3600 showed the best accuracy (p < 0.001). There were significant positive correlations for the volume and area of the abutments with accuracy (p < 0.001). The type, volume, and area of the clinically applicable abutments may affect the accuracy of intraoral scanners; however, the scanners used in the present study showed a clinically acceptable accuracy range, regardless of the type of abutment.