Scholarlink

<p><b>Background</b> The objective of this study was to develop a deep learning (DL) model for the detection and segmentation of six types of dental restorations and applications in panoramic radiographs of paediatric patients with mixed dentition.</p><p><b>Material and methods</b> A total of 2,033 panoramic radiographs were labelled for six different dental restorations. The dataset was divided into three parts: 80% for training, 10% for validation, and 10% for testing. The YOLOv8 model was trained for 500 epochs with a learning rate of 0.01. The success of the model was evaluated using sensitivity, precision and F1 score metrics.</p><p><b>Results</b> The YOLOv8 multiclass-DL model achieved high performance, with an overall F1 score of 0.89, supported by a sensitivity of 0.85 and precision of 0.93. Among the evaluated restoration types, dental fillings achieved the highest F1-score of 0.97, followed by stainless steel crowns with 0.94, space maintainers with 0.93, pulpotomies with 0.90, and root canal fillings with 0.84. The lowest performance was observed in the detection of dental brackets, which reached an F1-score of only 0.46.</p><p><b>Conclusion</b> YOLOv8-based DL models demonstrate a high level of success in detecting and segmenting dental restorations in panoramic radiographs of patients in the mixed dentition period.</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

Development of a YOLOv8-based deep learning model for detecting and segmenting dental restorations and dental applications in panoramic radiographs of mixed dentition

  • 期刊论文

  • 发表日期：

    2025年12月5日

查看全文
  • Alican Kuran,
  • Ibrahim Tevfik Gulsen,
  • Fatma Nur Kızılay,
  • Emine Gulsen,
  • Mustafa Enes Aşar,
  • Semanur Özüdoğru,
  • Türkan Ünal,
  • Ozer Celik,
  • İbrahim Şevki Bayrakdar,
  • Kaan Orhan

摘要

Background The objective of this study was to develop a deep learning (DL) model for the detection and segmentation of six types of dental restorations and applications in panoramic radiographs of paediatric patients with mixed dentition.

Material and methods A total of 2,033 panoramic radiographs were labelled for six different dental restorations. The dataset was divided into three parts: 80% for training, 10% for validation, and 10% for testing. The YOLOv8 model was trained for 500 epochs with a learning rate of 0.01. The success of the model was evaluated using sensitivity, precision and F1 score metrics.

Results The YOLOv8 multiclass-DL model achieved high performance, with an overall F1 score of 0.89, supported by a sensitivity of 0.85 and precision of 0.93. Among the evaluated restoration types, dental fillings achieved the highest F1-score of 0.97, followed by stainless steel crowns with 0.94, space maintainers with 0.93, pulpotomies with 0.90, and root canal fillings with 0.84. The lowest performance was observed in the detection of dental brackets, which reached an F1-score of only 0.46.

Conclusion YOLOv8-based DL models demonstrate a high level of success in detecting and segmenting dental restorations in panoramic radiographs of patients in the mixed dentition period.