Deep learning improves tooth disease detection from dental X-rays
Patient-Level, Leakage-Aware Deep Learning for Cross-Center Periapical Radiograph Classification
Computer Vision and Pattern Recognition
Summary
Detecting cavities and root problems using dental X-rays is important but hard to do automatically. The authors created a reliable way to test AI models that looks at images from different patients and clinics to avoid cheating. They tested different AI models and found that one called DenseNet121 works best across multiple clinics. Their method focuses on the right parts of the teeth, not on image quirks, making it trustworthy for future use.
What this means in practice
- •For medical ai developers: Train reliable tooth-level AI models to classify dental X-rays accurately across different clinics and patients.
- •For dental imaging device manufacturers: Incorporate validated AI benchmarks to improve automated disease detection features in dental X-ray equipment.$Commercial implications: Enables selling smarter dental X-ray machines with built-in diagnostic AI validated across multiple clinics.
Authors
Md Jubaer Rahman, Ulas Bagci
Abstract
Dental caries and endodontic disease are among the most common health conditions worldwide, and intraoral periapical radiographs are central to their detection, treatment planning, and follow-up. Automated tooth-level classification of these images, however, lacks reproducible benchmarks, is often evaluated with image-level splits that leak patients between training and test, and is rarely validated across clinics. This paper presents the first patient-level, leakage-aware classification benchmark for single-tooth intraoral periapical radiographs on the DentIRO dataset, which comprises 5,300 images from 3,243 patients across two clinics and four classes: Healthy, Caries, Crowned, and Root Canal. Five transfer-learning models are compared with patient-grouped stratified cross-validation, so that every patient remains within a single fold. DenseNet121 gave the strongest and most stable result at a mean macro-F1 of 0.9787, while the four ImageNet-initialized backbones performed comparably. A controlled comparison on a fixed architecture showed that chest-radiograph pretraining transferred less effectively than ImageNet initialization. Bidirectional cross-center validation produced a small average generalization gap of 0.0077, and Grad-CAM confirmed that predictions rely on clinically meaningful tooth regions rather than acquisition artifacts. The benchmark offers a rigorous and reproducible baseline for intraoral radiograph classification.