The Effect of Image Count on Accuracy in Digital Measurements in Dentistry.

Abstract

Objective: This study investigated how the number of images collected for digital measurements in dentistry affects accuracy compared with traditional methods. Methods: A Frasaco maxillary model was scanned using a SHINING 3D AutoScan-DS-MIX dental 3D scanner to create an STL file. The maxilla was molded 10 times using polyvinyl siloxane (Zhermack Elite HD+) to produce plaster models, which were scanned with the same reference scanner to generate 10 STL files. The Frasaco model was scanned 10 times, capturing images in intervals of 800-1000, 1000-1200, and 1200-1500 using a 3Shape TRIOS 3 intraoral scanner, creating additional STL files. These were analyzed with reverse engineering software. Results: The most accurate measurements were obtained using 1200-1500 images. Conventional impression techniques performed significantly worse. There was a significant difference between the groups Digital 1200-1500 and Plaster (p < 0.001) and between Digital 800-1000 and Plaster (p = 0.007). No significant difference was found when the digital groups were compared among themselves. There was also no significant difference between the Plaster and Digital 1000-1200 groups. To compare precision values that were normally distributed across three or more methods, a one-way ANOVA was used. Trueness values that were not normally distributed with three or more methods were compared employing the Kruskal-Wallis test. Conclusions: Different image counts affect digital measurement accuracy. The most accurate measurements were obtained when collecting 1200-1500 images. Conventional impression techniques were shown to perform significantly worse than digital impression.