Scholarlink

<p>Mandibular fractures are one of the most common injuries stemming from traumatic incidents. Proper and prompt diagnosis is vital to prevent permanent functional impairment and life-threatening complications. This study aims to develop an automated diagnostic tool to assist clinicians, using the latest deep learning architecture for the automatic detection and region-based classification of mandibular fractures. The deep learning architectures YOLOv8-seg and YOLOv8-cls were trained on a dataset of 330 and validated on 84 panoramic radiographs, also consisting of the less studied pediatric population. Our approach displayed a superior performance with an F1 score of 86%, surpassing the existing methods used for classifying mandibular fractures using panoramic radiographs. Furthermore, our proposed framework also effectively categorizes radiographs with plating/arch bars and mixed/permanent dentition, offering valuable support to healthcare professionals in the detection and classification of typical mandibular fractures.</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

Classification of mandibular fractures from panoramic radiographs using deep learning

  • 期刊论文

  • 发表日期：

    2025年4月16日

查看全文
  • Uttam Mittal,
  • Padmavati Khandnor,
  • Manoj Jaiswal,
  • Katinder Kaur,
  • Priyanka Rana,
  • Shiraz Mangat,
  • Harasees Singh,
  • Prakhar Sharma,
  • Balkaran Singh

摘要

Mandibular fractures are one of the most common injuries stemming from traumatic incidents. Proper and prompt diagnosis is vital to prevent permanent functional impairment and life-threatening complications. This study aims to develop an automated diagnostic tool to assist clinicians, using the latest deep learning architecture for the automatic detection and region-based classification of mandibular fractures. The deep learning architectures YOLOv8-seg and YOLOv8-cls were trained on a dataset of 330 and validated on 84 panoramic radiographs, also consisting of the less studied pediatric population. Our approach displayed a superior performance with an F1 score of 86%, surpassing the existing methods used for classifying mandibular fractures using panoramic radiographs. Furthermore, our proposed framework also effectively categorizes radiographs with plating/arch bars and mixed/permanent dentition, offering valuable support to healthcare professionals in the detection and classification of typical mandibular fractures.