Interpretable Attention-Guided Fused Dual-Model Features Framework for Enamel Caries Classification on Clinical Images

Dental enamel caries is among the most prevalent oral diseases worldwide. Early detection is essential, as incipient lesions can be managed with non-invasive therapies. A dataset of 2,000 clinical dental images categorized as Advanced Enamel Caries, Early-Stage Enamel Caries, and No Enamel Caries was curated and expanded to 12,000 images using preprocessing and augmentation. Two transfer learning models, Modified EfficientNetB0 and Modified MobileNetV2, were trained individually, then combined using an attention-guided fusion mechanism. Gradient-weighted Class Activation Mapping (Grad-CAM) was applied to provide visual interpretability. The Modified EfficientNetB0 and MobileNetV2 models achieved accuracies of 96.33% and 96.25%, respectively. The fused model with Random Forest demonstrated superior performance, achieving 96.92% accuracy, F1-score of 96.92 and ROC AUC of 99.34. Misclassifications were limited to adjacent disease stages, with no severe diagnostic errors.牙釉质龋是全球范围内最常见的口腔疾病之一。早期检测至关重要，因为早期龋损可通过非侵入性疗法进行干预管理。本研究遴选了2000张临床牙科图像构建数据集，这些图像被划分为重度牙釉质龋、早期牙釉质龋以及无牙釉质龋三个类别；随后通过预处理与数据增强技术将数据集扩充至12000张。分别训练了改进型EfficientNetB0与改进型MobileNetV2两个迁移学习模型，随后通过注意力引导融合机制将二者进行结合。采用梯度加权类激活映射（Gradient-weighted Class Activation Mapping, Grad-CAM）技术以提供可视化可解释性。 改进型EfficientNetB0与改进型MobileNetV2模型的分类准确率分别为96.33%与96.25%。结合随机森林的融合模型表现更优，其分类准确率达96.92%，F1值为96.92，受试者工作特征曲线下面积（ROC AUC）为99.34。分类错误仅局限于相邻疾病阶段之间，未出现严重的诊断失误。