Related Experiment Video
Updated: Jul 24, 2025
Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
Published on: August 5, 2021
Age estimation based on 3D pulp segmentation of first molars from CBCT images using U-Net
Yangjing Song1, Huifang Yang2, Zhipu Ge3
1Department of Oral and Maxillofacial Radiology, Peking University School and Hospital of Stomatology; National Center of Stomatology & National Clinical Research Center for Oral Diseases & National Engineering Laboratory for Digital and Material Technology of Stomatology & Beijing Key Laboratory of Digital Stomatology & Research Center of Engineering and Technology for Computerized Dentistry Ministry of Health & NMPA Key Laboratory for Dental Materials, Beijing, China.
A U-Net model accurately segments first molar pulp cavities from cone-beam CT scans. The resulting pulp cavity volumes enable reliable human age estimation with good precision and accuracy.
Area of Science:
- Forensic dentistry
- Radiology
- Artificial intelligence in medicine
Background:
- Accurate age estimation is crucial in forensic and clinical contexts.
- Dental age assessment methods often rely on subjective interpretations or invasive procedures.
- Cone-beam computed tomography (CBCT) provides detailed 3D imaging of dental structures.
Purpose of the Study:
- To develop a U-Net model for precise segmentation of the intact pulp cavity in first molars.
- To establish a mathematical model for human age estimation based on pulp cavity volume.
- To evaluate the accuracy and precision of the developed age estimation model.
Main Methods:
- A U-Net model was trained using 20 sets of CBCT images to segment first molar pulp cavities.
- Pulp cavity volumes were calculated for 239 maxillary and 234 mandibular first molars from individuals aged 15-69.
- Logarithmic regression analysis was performed to create an age estimation model, with subsequent validation on 256 additional first molars.
Main Results:
- The U-Net model achieved a high segmentation accuracy with a Dice similarity coefficient of 95.6%.
- The established age estimation model demonstrated a coefficient of determination (R²) of 0.662.
- The model yielded a mean absolute error of 6.72 years and a root mean square error of 8.26 years.
Conclusions:
- The trained U-Net model accurately segments first molar pulp cavities from 3D CBCT images.
- The segmented pulp cavity volumes provide a reliable basis for estimating human age.
- This AI-driven approach offers a precise and accurate method for dental age assessment.
