Accuracy of a Single Molar Crown Designed Using Artificial Intelligence and Computer-aided Design Software: An In Vitro Study

Abstract

Purpose: The purpose of this study was to compare the geometric accuracy of single molar crown created by artificial intelligence (AI) and morphology generated by postgraduate students, faculty members, and dental technicians using CAD software. Materials and methods: An AI model called CrownGen was developed using 200 digital scans of maxillary arches. Each scan was duplicated into two identical sets. One set contained the original maxillary arch. In the second set, the right maxillary first molar was digitally removed using Autodesk Meshmixer. These paired datasets (original and modified) were used to train a generative adversarial network (GAN) to recreate missing teeth. To test CrownGen’s performance, a comparative study was carried out with 88 samples. The right maxillary first molar in each case was prepared for an all-ceramic crown, and intraoral scans were taken. Four groups designed crowns using CAD software (n = 22 each): Group I—CrownGen (AI software), group II—postgraduate students, group III—faculty, group IV—dental technicians. Mean Hausdorff distance (HD) was used for assessing surface deviation and morphological accuracy. A statistical analysis was performed using one-way analysis of variance (ANOVA) and Tukey’s post hoc test. Results: The mean HD values showed a statistically significant difference between all the four groups (p < 0.001). The maximum mean HD value was recorded for group IV (technicians) at 0.582 mm, while the minimum mean HD value was recorded for group I (AI–CrownGen) at 0.368 mm. AI-generated (CrownGen) crowns showed the closest geometric resemblance to the reference tooth, followed by postgraduate and faculty designs. Technician-designed crowns exhibited the greatest deviation. Conclusion: AI-based crown design (CrownGen) demonstrates high accuracy in replicating the original tooth morphology in comparison to crowns designed using human intervention. Deviations in faculty and technician designs may reflect clinically intentional adjustments. These findings support the integration of AI in dental crown design while emphasizing the need for human oversight in complex clinical cases.