Smart Dental Caries Detection System

Abstract

DentAI is an intelligent dental screening application designed to detect dental caries (tooth decay) from panoramic X-ray images using deep learning techniques. It utilizes a Convolutional Neural Network (CNN) model trained on dental radiographs to identify radiolucent regions and structural anomalies associated with caries. Built with Streamlit, the system provides an interactive interface where users can upload X-ray images and receive real-time analysis. The preprocessing pipeline standardizes image size and normalizes pixel values before prediction, and the model classifies results into three categories: caries detected, no caries detected, and uncertain, with a confidence score to improve reliability. The application also enhances interpretability by offering detailed explanations and integrates an AI-powered chatbot to answer dental health queries and provide guidance based on the results. DentAI aims to support early detection and preliminary screening, particularly in resource-limited environments, while emphasizing that it is not a substitute for professional dental diagnosis.

Introduction

Dental caries (tooth decay) is a widespread oral health issue, and early detection is crucial to prevent complications. Traditional detection relies on manual examination of panoramic X-rays, which can be time-consuming and inconsistent.

The DentAI project uses Convolutional Neural Networks (CNNs) to automate caries detection from panoramic dental X-ray images. Users upload images through a Streamlit-based web interface, and the system provides real-time predictions with confidence scores and explanations.

Key features and methodology:

• Image preprocessing: Conversion to RGB, normalization, resizing to 256×256 pixels.
• CNN detection: Extracts features like radiolucent regions and enamel degradation to classify caries presence.
• Web-based implementation: Python, TensorFlow/Keras for the model, OpenCV/PIL for image processing, Streamlit for the UI.
• Chatbot integration: AI-powered assistant provides additional guidance and answers dental health queries.

Experimental results:

• The system achieved ~94% accuracy in detecting caries.
• Introduces an uncertain category (0.40–0.60 probability) to handle borderline cases.
• Provides near real-time results and interpretable outputs highlighting detected anomalies.

Conclusion

The proposed DentAI system demonstrates the effective use of deep learning techniques for automated dental caries detection from panoramic X-ray images. By integrating a Convolutional Neural Network (CNN) with a user-friendly Streamlit interface, the system provides real-time analysis, accurate predictions, and explainable results. The introduction of a threshold-based classification, including an uncertain category, enhances reliability and reduces the chances of misclassification. Additionally, the inclusion of an AI-powered chatbot improves user interaction by offering guidance and answering dental health-related queries. Despite its promising performance, the system has certain limitations, such as dependency on image quality and the use of a relatively limited dataset. In future work, the model can be improved by training on larger and more diverse datasets to enhance generalization. Advanced techniques such as transfer learning, attention mechanisms, and segmentation- based approaches can be incorporated to improve detection accuracy, especially for early-stage caries. Furthermore, integration with real clinical systems and mobile applications can increase accessibility and usability. The system can also be extended to detect other dental conditions such as periodontal diseases and lesions. In conclusion, DentAI serves as an efficient and intelligent assistive tool for preliminary dental screening, supporting early diagnosis while emphasizing that it is not a replacement for professional dental expertise.

References

[1] A. Esteva, B. Kuprel, R. A. Novoa et al., “Dermatologist-level classification of skin cancer with deep neural networks,” Nature, vol. 542, no. 7639, pp. 115–118, 2017. [2] D. C. Cire?an, U. Meier, J. Schmidhuber, “Multi- column deep neural networks for image classification,” Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), pp. 3642–3649, 2012. [3] J. Lee, S. Kim, E. Jeong, and H. Choi, “Detection and diagnosis of dental caries using a deep learning- based convolutional neural network algorithm,” Journal of Dentistry, vol. 77, pp. 106–111, 2018. [4] M. Moutselos, I. Maglogiannis, “Deep learning approaches for dental image analysis: A review,” Artificial Intelligence in Medicine, vol. 107, 2020. [5] S. Litjens, T. Kooi, B. E. Bejnordi et al., “A survey on deep learning in medical image analysis,” Medical Image Analysis, vol. 42, pp. 60–88, 2017. [6] K. Simonyan, A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” International Conference on Learning Representations (ICLR), 2015. [7] F. Chollet, “Xception: Deep learning with depthwise separable convolutions,” Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), pp. 1251–1258, 2017. [8] T. F. Chan, L. A. Vese, “Active contours without edges,” IEEE Transactions on Image Processing, vol. 10, no. 2, pp. 266–277, 2001. [9] Streamlit Inc., “Streamlit: The fastest way to build and share data apps,” [Online]. Available: https://streamlit.io [10] TensorFlow, “An end-to-end open-source machine learning platform,” [Online].

Copyright

Copyright © 2026 Srinivasan S, Amuthapriyan A M, Andrew Jebish V, Thudhin N S. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.