Related Experiment Video
Updated: Jan 14, 2026
05:49
Author Spotlight: Advancing CBCT and Digital Dental Image Integration with AI-Assisted Digitization
Published on: February 23, 2024
1.3K
Deep learning-based assessment of periapical radiographic image quality
Xiuting Chi1, Mingchao Wang2, Yue Gao1
1Department of Radiology, Qingdao Stomatological Hospital Affiliated to Qingdao University, Qingdao, 26001, Shandong, China.
Scientific Reports
|January 12, 2026
Summary
This study introduces an AI system for automated dental X-ray (periapical radiograph) quality assessment, improving accuracy and reducing patient radiation exposure from retakes.
Area of Science:
- Artificial Intelligence in Dentistry
- Medical Imaging Analysis
- Deep Learning for Quality Assessment
Background:
- Manual assessment of periapical radiographs (PAs) is subjective and inefficient.
- Current methods lead to diagnostic uncertainty and increased patient radiation.
- Need for automated, objective quality control in dental imaging.
Purpose of the Study:
- Develop an automated deep learning system for comprehensive PA quality assessment.
- Classify tooth positions and detect common image quality defects.
- Enhance diagnostic accuracy and workflow efficiency in dental clinics.
Main Methods:
- Utilized a retrospective dataset of 3594 PAs.
- Trained ResNet50 models for tooth position classification and defect detection.
- Employed data augmentation and oversampling to address class imbalance.
Main Results:
- Achieved high performance with AUC values up to 1.000 for various assessments.
- Demonstrated excellent accuracy in identifying tooth positions and quality defects.
- AUCs: 0.997 (tooth position), 0.996 (vertical angle), 1.000 (horizontal angle), 1.000 (crown coverage), 0.994 (apical coverage), 0.999 (cone cut), 0.924 (scratch).
Conclusions:
- The ResNet50 algorithm effectively automates PA quality assessment.
- AI tool shows potential for clinical integration to improve diagnostics and reduce retakes.
- Further validation on multi-center datasets is needed before clinical deployment.
