Scholarlink

Objective <p>To compare tuned end-to-end and hybrid deep learning strategies for image-based classification of common oral conditions under small and imbalanced data conditions, and to determine whether the same pattern persists in a clinically focused periodontal subset. </p> Materials and methods <p>A total of 10,735 oral images were analyzed. We compared tuned end-to-end and frozen-feature hybrid models across ResNet18, EfficientNetV2-S, and ConvNeXt-Tiny backbones. Post-hoc classifiers included K-nearest neighbors (KNN), random forest (RF), linear support vector machine (SVM), and radial basis function (RBF)-SVM. In addition to the original holdout setting, stratified 5-fold cross-validation, class-imbalance-oriented ablations, paired McNemar tests, Grad-CAM, and ROC/PR analyses were performed. A periodontal 3-class subtask (Calculus/Caries/Gingivitis) was also evaluated.</p> Results <p>On the full holdout task, the strongest end-to-end model was ConvNeXt-Tiny at 224 × 224 (accuracy 0.9322, macro-F1 0.9177), whereas the strongest overall holdout model was ConvNeXt-Tiny + RBF-SVM (accuracy 0.9396, macro-F1 0.9242). On the periodontal holdout task, the strongest model was ConvNeXt-Tiny + linear SVM (accuracy 0.8889, macro-F1 0.8727). Under pooled 5-fold cross-validation, hybrid gains remained statistically supported for several backbone-specific comparisons, including full-task ResNet18 end-to-end versus RBF-SVM (macro-F1 0.8984 vs. 0.9130; McNemar <i>p</i> = 5.41e-08). The dominant error pattern remained mutual confusion between Calculus and Gingivitis.</p> Conclusion <p>In the task of oral image classification, the hybrid model strategy demonstrates improved performance metrics over standard end-to-end models, particularly within the clinically focused periodontal subset. While further refinement is needed to address specific overlaps between closely related classes (such as Calculus and Gingivitis), this approach proves to be an effective optimization strategy. Ultimately, it provides valuable methodological options and foundational insights for the future development of intelligent periodontal auxiliary diagnostic tools.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Research on the efficacy of hybrid deep learning models for image-based classification of common oral conditions

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月11日

查看全文
  • Yiqian Xia,
  • Weisong Liu,
  • Zhongda Liu,
  • Bo Lan

摘要

Objective

To compare tuned end-to-end and hybrid deep learning strategies for image-based classification of common oral conditions under small and imbalanced data conditions, and to determine whether the same pattern persists in a clinically focused periodontal subset.

Materials and methods

A total of 10,735 oral images were analyzed. We compared tuned end-to-end and frozen-feature hybrid models across ResNet18, EfficientNetV2-S, and ConvNeXt-Tiny backbones. Post-hoc classifiers included K-nearest neighbors (KNN), random forest (RF), linear support vector machine (SVM), and radial basis function (RBF)-SVM. In addition to the original holdout setting, stratified 5-fold cross-validation, class-imbalance-oriented ablations, paired McNemar tests, Grad-CAM, and ROC/PR analyses were performed. A periodontal 3-class subtask (Calculus/Caries/Gingivitis) was also evaluated.

Results

On the full holdout task, the strongest end-to-end model was ConvNeXt-Tiny at 224 × 224 (accuracy 0.9322, macro-F1 0.9177), whereas the strongest overall holdout model was ConvNeXt-Tiny + RBF-SVM (accuracy 0.9396, macro-F1 0.9242). On the periodontal holdout task, the strongest model was ConvNeXt-Tiny + linear SVM (accuracy 0.8889, macro-F1 0.8727). Under pooled 5-fold cross-validation, hybrid gains remained statistically supported for several backbone-specific comparisons, including full-task ResNet18 end-to-end versus RBF-SVM (macro-F1 0.8984 vs. 0.9130; McNemar p = 5.41e-08). The dominant error pattern remained mutual confusion between Calculus and Gingivitis.

Conclusion

In the task of oral image classification, the hybrid model strategy demonstrates improved performance metrics over standard end-to-end models, particularly within the clinically focused periodontal subset. While further refinement is needed to address specific overlaps between closely related classes (such as Calculus and Gingivitis), this approach proves to be an effective optimization strategy. Ultimately, it provides valuable methodological options and foundational insights for the future development of intelligent periodontal auxiliary diagnostic tools.