Scholarlink

The identification of dental carries is crucial for prompt diagnosis and intervention in clinical dental practice. This study concentrates on the creation of a deep learning model aimed at the automatic identification of dental caries through the analysis of radiographic bitewing images. The model utilizes the U-Net architecture with ResNet backbone to capitalize on the benefits of transfer learning, especially considering the limited size of the dataset accessible. This paper demonstrates the efficiency of U-Net combined with transfer learning and CNN model in detection of dental carries using a augmented dataset of 300 clinically collected radiographic bitewing images. The proposed methodology is capable of significantly enhancing the diagnostic ability of dental professionals and enabling quicker and more accurate decisions in the clinical setting.

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

Auto Segmentation and Classification for Dental Carries Detection Using U-Net and CNN Models

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • K. S. Ravi Narayana,
  • P. K. Uma,
  • P. Govindaraj,
  • A. Divya

摘要

The identification of dental carries is crucial for prompt diagnosis and intervention in clinical dental practice. This study concentrates on the creation of a deep learning model aimed at the automatic identification of dental caries through the analysis of radiographic bitewing images. The model utilizes the U-Net architecture with ResNet backbone to capitalize on the benefits of transfer learning, especially considering the limited size of the dataset accessible. This paper demonstrates the efficiency of U-Net combined with transfer learning and CNN model in detection of dental carries using a augmented dataset of 300 clinically collected radiographic bitewing images. The proposed methodology is capable of significantly enhancing the diagnostic ability of dental professionals and enabling quicker and more accurate decisions in the clinical setting.