الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 133 |  |  |
| | | from | 133 | | |
Abstract
Estimating chronological age is a crucial task with significant importance in various fields, including forensics, clinical procedures, illegal immigration, criminal investigation, and civil issues. Age estimation can be accomplished through different biometric methods such as facial analysis, tooth examination, bone analysis, and voice analysis. In this thesis, the focus is specifically on age estimation using teeth due to their reliable nature as age indicators, given their development at different stages of life. The manual methods for dental age estimation can be intricate, particularly when dealing with adults and seniors. Therefore, this study aimed to address this challenge by conducting a comparative survey of convolutional neural network models used for dental age estimation, specifically from radiological images like cone-beam computed tomography (CBCT), magnetic resonance imaging (MRI), and orthopantomography (OPG). Additionally, we proposed a new CNN architecture, referred to as the custom CNN model, that aimed to automatically estimate dental age based on dental radiographs (OPG) from Egyptian populations. A total of 1,007 images of male and female subjects aged 6 to 70 years were collected for the study. The performance of the custom model was compared against two pre-trained (VGG-19, Xception) models commonly used in computer vision tasks. The results of the study revealed that the custom CNN model outperformed the pre-trained models, achieving an average accuracy of 94% in automatically predicting dental age across all age groups.