Scholarlink

Background <p>Evaluating the quality of root canal filling (RCF) performed by dental students in preclinical settings is a time-consuming process for clinicians and is often subjectively assessed.</p> Methods <p>This study proposes deep learning (DL)-based RCFLA-YOLO model for the objective and automatic evaluation of RCF length (RCFL), one of the key parameters in assessing RCF quality from periapical radiographs (PRs). To develop the proposed method, an RCFL dataset comprising 735 PRs labeled by expert clinicians was created. The method uses YOLOv11 architectures, one of the state-of-the-art DL-based object detection techniques, and its performance has been validated with this dataset.</p> Results <p>Performance evaluations conducted on the RCFL dataset demonstrated that among the tested models, the YOLOv11m architecture achieved the highest performance, with 77.51% precision, 79.03% recall, 78.28% F1-score, and 87.89% mAP50.</p> Conclusions <p>A performance comparison with similar studies in the literature indicates that the proposed method is among the top-performing approaches based on the applied evaluation metrics. Furthermore, this study has the potential to be one of the first to assess RCF quality in student procedures. In conclusion, the proposed method, with its notable strong performance, demonstrates its potential as an effective tool for developing decision support systems aimed at the automatic evaluation of RCF quality in preclinical dental education.</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

RCFLA-YOLO: a deep learning-driven framework for the automated assessment of root canal filling quality in periapical radiographs

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月1日

查看全文
  • Muhammed Ayhan,
  • İsmail Kayadibi,
  • Berkehan Aykanat

摘要

Background

Evaluating the quality of root canal filling (RCF) performed by dental students in preclinical settings is a time-consuming process for clinicians and is often subjectively assessed.

Methods

This study proposes deep learning (DL)-based RCFLA-YOLO model for the objective and automatic evaluation of RCF length (RCFL), one of the key parameters in assessing RCF quality from periapical radiographs (PRs). To develop the proposed method, an RCFL dataset comprising 735 PRs labeled by expert clinicians was created. The method uses YOLOv11 architectures, one of the state-of-the-art DL-based object detection techniques, and its performance has been validated with this dataset.

Results

Performance evaluations conducted on the RCFL dataset demonstrated that among the tested models, the YOLOv11m architecture achieved the highest performance, with 77.51% precision, 79.03% recall, 78.28% F1-score, and 87.89% mAP50.

Conclusions

A performance comparison with similar studies in the literature indicates that the proposed method is among the top-performing approaches based on the applied evaluation metrics. Furthermore, this study has the potential to be one of the first to assess RCF quality in student procedures. In conclusion, the proposed method, with its notable strong performance, demonstrates its potential as an effective tool for developing decision support systems aimed at the automatic evaluation of RCF quality in preclinical dental education.