Scholarlink

<p>Dental characteristics are increasingly considered valuable indicators for adult age estimation in forensic and clinical contexts, yet conventional methods struggle with complex dental data. Conventional linear regression relies on predefined equations and the number of dental characteristics aggregated across the dentition, which may lead to information loss and reduced accuracy. This study investigated machine learning methods to enhance age estimation using encoded dental characteristics. Panoramic radiographs of 2,415 individuals aged 20 to 89 years were analyzed, and dental characteristics were encoded using a modified one-hot encoding approach that preserved dual-feature information. A numerical study compared linear regression and six machine learning methods, including random forest, extreme gradient boosting, k-nearest neighbors, support vector machine, decision tree, and multi-layer perceptron. Model performance was evaluated using the root mean squared error (RMSE) and R-squared statistic on the test dataset. In the sex-aggregated dataset, linear regression achieved an RMSE of 12.04 years with an R-squared of 0.63, whereas the best-performing machine learning model achieved an RMSE of 10.88 years with an R-squared of 0.70. SHapley Additive exPlanations (SHAP) analysis revealed that sound and missing teeth were the most influential predictors of age estimation. Given the magnitude of prediction error, the proposed models should be interpreted as adjunctive tools for dental age estimation in adults.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine learning approach for adult age estimation using dental characteristics on panoramic radiographs

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月24日

查看全文
  • Donghwan Lee,
  • Sehyun Oh,
  • Seyeon Hwang,
  • Jiae Park,
  • Yujin Oh,
  • Akiko Kumagai,
  • Sang-Seob Lee

摘要

Dental characteristics are increasingly considered valuable indicators for adult age estimation in forensic and clinical contexts, yet conventional methods struggle with complex dental data. Conventional linear regression relies on predefined equations and the number of dental characteristics aggregated across the dentition, which may lead to information loss and reduced accuracy. This study investigated machine learning methods to enhance age estimation using encoded dental characteristics. Panoramic radiographs of 2,415 individuals aged 20 to 89 years were analyzed, and dental characteristics were encoded using a modified one-hot encoding approach that preserved dual-feature information. A numerical study compared linear regression and six machine learning methods, including random forest, extreme gradient boosting, k-nearest neighbors, support vector machine, decision tree, and multi-layer perceptron. Model performance was evaluated using the root mean squared error (RMSE) and R-squared statistic on the test dataset. In the sex-aggregated dataset, linear regression achieved an RMSE of 12.04 years with an R-squared of 0.63, whereas the best-performing machine learning model achieved an RMSE of 10.88 years with an R-squared of 0.70. SHapley Additive exPlanations (SHAP) analysis revealed that sound and missing teeth were the most influential predictors of age estimation. Given the magnitude of prediction error, the proposed models should be interpreted as adjunctive tools for dental age estimation in adults.