An Automated Cavity Level Prediction based on Dental Imaging Sensors by using Enhanced AI Assisted Learning Principles

Dental cavities are one of the most prevalent oral health issues affecting individuals worldwide. Early detection and treatment of cavities are essential to prevent severe dental diseases that can lead to tooth loss and other complications. This paper presents an automated cavity level prediction system based on dental X-ray images using a hybrid model that integrates ResNet-50 for deep feature extraction and XGBoost for classification. The proposed model addresses the challenges of accurate cavity detection, leveraging state-of-the-art techniques in deep learning and boosting algorithms. The dataset comprises high-resolution dental X-rays, annotated with different cavity levels (mild, moderate, severe, and healthy teeth), to train and evaluate the model. A comprehensive evaluation was conducted, comparing the proposed model’s performance with nine existing machine learning models, including CNN, SVM, Random Forest, and Gradient Boosting. The results demonstrate that the proposed model significantly outperforms all other models, achieving an accuracy of $\mathbf{9 7. 8 1 \%}$, a precision of $\mathbf{9 6. 5 4 \%}$, a recall of $\mathbf{9 7. 2 3 \%}$, and an F1-score of $\mathbf{9 6. 8 8 \%}$. Additionally, the model exhibits competitive training time and memory usage, making it suitable for real-time dental diagnostic applications. Robustness was validated through $\mathbf{5}$-fold cross-validation and AUC-ROC analysis, further establishing the proposed model’s reliability in cavity detection. These findings highlight the effectiveness of integrating deep learning and boosting for medical imaging tasks and present a viable solution for automated dental diagnostics.