Elbashti Mahmoud E, Paz-Cortes Marta Macarena, Giovannini Giovanni, Acero-Sanz Julio, Abou-Ayash Samir, Çakmak Gülce, Molinero-Mourelle Pedro
Department of Advanced Prosthodontics, Graduate School of Medical and Dental Sciences, Institute of Science Tokyo, Tokyo, Japan and Faculty of Dentistry, University of Zawia, Zawia, Libya.
Faculty of Dentistry, Alfonso X el Sabio University, Madrid, Spain.
J Dent. 2025 May;156:105652. doi: 10.1016/j.jdent.2025.105652. Epub 2025 Feb 25.
To develop a scripting-based technique for managing three-dimensional (3D) dental data and evaluate the regenerated standard tessellation language (STL) data in terms of file size, accuracy (trueness and precision), and processing time.
Ten STL dental and maxillofacial models were obtained from various imaging technologies, including intraoral scanners, computer-aided design (CAD) software, and cone-beam computed tomography (CBCT), and saved as STL files. ChatGPT was used to generate Python scripts in Blender for mesh simplification and data compression, which were then saved as .py files. The models were regenerated from these scripts in Blender, and their accuracy was assessed using GOM Inspect software, comparing trueness and precision. Statistical analysis, including Kruskal-Wallis and Mann-Whitney tests, was conducted to evaluate differences in file sizes between the original, Python-generated, and regenerated STL files, with statistical analyses performed at a level of significance α=0.05.
The scripting-based technique was successfully utilized in ChatGPT to generate Python script code for accessing comprehensive data on STL models, utilizing Blender's scripting functionality. This approach enabled the generation, regeneration, and visualization of STL models, resulting in significantly smaller file sizes for both the Python script and regenerated STL files compared to the original STL files (p < 0.001). No significant differences in trueness were observed, with deviations ranging from 0.0 µm to 6.8 µm, and all regenerated STL models demonstrated perfect precision. Additionally, a proportional relationship was noted between the original STL file sizes and processing times.
The scripting-based approach proved to be effective in coding, storing, and regenerating STL dental data with reduced file sizes and efficient processing times without compromising the accuracy.
Various STL dental models of patients can be coded, stored, and regenerated to be used again within efficient processing time without affecting the accuracy.
开发一种基于脚本的技术来管理三维（3D）牙科数据，并从文件大小、准确性（真实性和精确性）和处理时间方面评估重新生成的标准镶嵌语言（STL）数据。
从包括口腔内扫描仪、计算机辅助设计（CAD）软件和锥形束计算机断层扫描（CBCT）在内的各种成像技术中获取了10个STL牙科和颌面模型，并保存为STL文件。使用ChatGPT在Blender中生成用于网格简化和数据压缩的Python脚本，然后将其保存为.py文件。在Blender中根据这些脚本重新生成模型，并使用GOM Inspect软件评估其准确性，比较真实性和精确性。进行了包括Kruskal-Wallis和Mann-Whitney检验在内的统计分析，以评估原始、Python生成和重新生成的STL文件之间的文件大小差异，统计分析在显著性水平α=0.05下进行。
基于脚本的技术在ChatGPT中成功用于生成Python脚本代码，以利用Blender的脚本功能访问STL模型的全面数据。这种方法实现了STL模型的生成、重新生成和可视化，与原始STL文件相比，Python脚本和重新生成的STL文件的文件大小显著更小（p < 0.001）。在真实性方面未观察到显著差异，偏差范围为0.0 µm至6.8 µm，所有重新生成的STL模型都显示出完美的精确性。此外，注意到原始STL文件大小与处理时间之间存在比例关系。
基于脚本的方法被证明在编码、存储和重新生成STL牙科数据方面是有效的，文件大小减小且处理时间高效，同时不影响准确性。
患者的各种STL牙科模型可以进行编码、存储和重新生成，以便在高效的处理时间内再次使用而不影响准确性。
