DAPR: Dynamic Distribution-Aware and Adaptive Pseudo-Label Refinement for Long-Tailed Semi-Supervised Oral Disease Classification
Open Access
- 15 August 2026
- journal article
- research article
- Published by MDPI AG in Electronics
- Vol. 15 (16) , 3641
Abstract
Oral disease image classification can support computer-assisted assessment of intraoral images. However, obtaining large-scale annotated medical data is expensive, while real-world oral disease datasets often exhibit severe long-tailed distributions, where minority disease categories contain only limited samples. Existing semi-supervised learning methods commonly rely on fixed-threshold pseudo labels and may produce prediction distributions dominated by majority classes, resulting in class bias and pseudo-label noise accumulation under long-tailed settings. To address these issues, we propose DAPR, a Distribution-Aware and Adaptive Pseudo-Label Refinement framework for long-tailed semi-supervised oral disease classification. DAPR employs Dynamic Category Distribution Modeling (DCDM) to track the evolving prediction distribution of unlabeled samples and generate distribution-aware soft pseudo labels. A class-adaptive dynamic thresholding (CADT) mechanism was further introduced to improve minority-class sample utilization. In addition, Relation-aware Representation Learning (RRL) aligns semantic and feature relationships to enhance feature discrimination. Experiments using stratified five-fold cross-validation on a long-tailed oral disease dataset show that DAPR achieves the highest average Accuracy and Macro-F1 among the compared methods under the evaluated setting. DAPR achieves the highest average Accuracy and Macro-F1 among the compared methods and obtains strong aggregate tail-class performance, particularly for Tooth Discoloration and Ulcers. These results indicate that DAPR improves aggregate class-balanced learning under the evaluated dataset, annotation ratio, and backbone configuration.Keywords
Funding Information
- National Natural Science Foundation of China (62301473)
- Jiangsu University Qing Lan Project
- Funding for School-Level Research Projects of Yancheng Institute of Technology (xjr2023012)
- Yancheng Basic Research Fund Project (YCBK2024028)
- Postgraduate Research & Practice Innovation Program of Jiangsu Province (SJCX25_2203)
This publication has 19 references indexed in Scilit:
- A review of ultrasound video segmentation: From temporal modeling to clinical utilityNeurocomputing, 2026
- Enhancing long-range contextual interactions via frequency-aware directional shifts for medical image segmentationApplied Soft Computing, 2026
- A Systematic Review on Long-Tailed LearningIEEE Transactions on Neural Networks and Learning Systems, 2025
- Diagnosing oral and maxillofacial diseases using deep learningScientific Reports, 2024
- Annotation-efficient deep learning for automatic medical image segmentationNature Communications, 2021
- Guest Editorial Annotation-Efficient Deep Learning: The Holy Grail of Medical ImagingIEEE Transactions on Medical Imaging, 2021
- Annotation-Efficient Learning for Medical Image Segmentation Based on Noisy Pseudo Labels and Adversarial LearningIEEE Transactions on Medical Imaging, 2020
- Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based LocalizationInternational Journal of Computer Vision, 2019
- Oral diseases: a global public health challengeThe Lancet, 2019
- A survey on deep learning in medical image analysisMedical Image Analysis, 2017