article · Dentistry Journal
Background: Artificial intelligence (AI) systems are increasingly being used in digital smile design and esthetic treatment planning; however, evidence comparing the esthetic performance of AI-generated designs with expert clinician-generated designs remains limited. Objective evaluation using standardized esthetic indices is necessary to determine whether AI-generated outcomes achieve comparable clinical quality. Methods: This prospective paired comparative study included 33 patients. For each case, two smile designs were created: one generated using a fully automated AI system (SmileFy) and one designed manually by an experienced clinician using Exocad software(version 3.2 Elefsina; exocad GmbH, Darmstadt, Germany). Twenty blinded prosthodontists evaluated all designs using the Dental Esthetic Screening Index (DESI) and a visual analog scale (VAS). Patients provided esthetic VAS ratings and forced-choice preferences. Objective geometric measurements and total design time were recorded. Paired statistical analyses were performed with a significance level of p < 0.05. Results: AI-generated designs demonstrated significantly lower total DESI scores than expert-generated designs (14.79 ± 1.63 vs. 18.73 ± 1.82; p < 0.001). Both expert and patient VAS ratings were significantly higher for AI designs (p < 0.001). Patients preferred AI-generated designs in 69.7% of cases compared with 30.3% for expert designs (p < 0.001). AI workflows were significantly faster, with a mean design time of 30.82 ± 5.14 min versus 63.48 ± 14.12 min for expert workflows, corresponding to a 51.46% reduction in planning time (p < 0.001). Conclusions: Fully automated AI-generated smile designs demonstrated favorable esthetic performance, higher patient acceptance, and substantial improvements in workflow efficiency compared with expert-driven digital designs, supporting their potential role as adjunctive tools in esthetic treatment planning.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.3390/dj14030166
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
