Related Experiment Video
Updated: Jul 19, 2025
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
2.5K
Denoised encoder-based residual U-net for precise teeth image segmentation and damage prediction on panoramic
Sultan A Almalki1, Shtwai Alsubai2, Abdullah Alqahtani3
1Department of Preventive Dental Sciences, College of Dentistry, Prince Sattam Bin AbdulAziz University, Al-Kharj 11942, Saudi Arabia.
Journal of Dentistry
|August 8, 2023
Summary
This study introduces a robust AI model for precise teeth segmentation in dental radiographs. The denoised U-Net model accurately identifies dental issues, improving diagnostic capabilities.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer Vision
Background:
- Accurate teeth segmentation is crucial for diagnosing dental conditions.
- Existing methods may struggle with complex dental fillings and varied tooth types.
Purpose of the Study:
- To develop and evaluate a robust AI model for precise teeth segmentation in panoramic radiographs.
- To enhance the adaptability and accuracy of segmentation models for diverse dental datasets.
Main Methods:
- Utilized a denoised encoder-based residual U-Net model for teeth segmentation.
- Pre-processed the Tufts dataset, resizing images and employing a modified identity block for finer segmentation.
- Incorporated a denoised block to effectively handle noisy ground truth images.
Main Results:
- The proposed model achieved high performance metrics, with mean Dice and mean IoU scores of 98.90% and 98.74%, respectively.
- Demonstrated precise teeth segmentation on the Tufts dental dataset, even with dense dental fillings and various tooth types.
Conclusions:
- The AI-enabled model offers a precise approach to segment teeth in dental radiographs.
- This technology has the potential to significantly improve dental diagnostics and patient care.
