EUVIMEDEUVIMED entdecken
Crossref

Difficulty-Aware Pseudo-Label Correction Network for Fine-Grained Classification of Choroidal Neovascularization in OCT Images

Lingzhao Meng, Xiaoming Xi, Jia Han, Lishan Qiao, Yilong Yin, Xinjian Chen

IEEE Transactions on Multimedia · 2026 · 28 · 7163-7175

DOI: 10.1109/tmm.2026.3673510 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Multimedia, 2026. Verfasst von Lingzhao Meng und 5 weiteren Autor:innen. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 43 Literaturangaben. Gefördert von: Key R&D Program of Shandong Province, National Natural Science Foundation of China, Key Technology Research and Development Program of Shandong Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Choroidal neovascularization (CNV) classification is a fine-grained classification task. Accurate classification of CNV in optical coherence tomography (OCT) images is crucial for clinical treatment. However, image acquisition noise degrades image quality and exacerbates confirmation bias from class imbalance in medical datasets. Moreover, significant inter-class ambiguity in fine-grained categories can misclassify informative samples (e.g., hard samples or minority class samples) when generating pseudo-labels, leading to sub-optimal classifiers. To address these challenges, we propose a difficulty-aware pseudo-label correction network (DPLC-Net). Specifically, we designed a robust feature mining module using feature similarity loss to maintain consistency between generated adversarial and original samples, enabling noise-resistant feature learning. A difficulty-aware pseudo-label correction module mines and corrects potential noisy pseudo-labels to improve classification performance. Finally, to alleviate data bias and leverage all unlabeled samples, we integrated a hybrid consistency and pseudo-labeling module comprising adaptive weighted consistency loss (AWCL) and class-aware dynamic threshold strategy (CDTS). AWCL adaptively learns weights for unlabeled samples, effectively utilizing all unlabeled data through weighted consistency loss. CDTS dynamically adjusts confidence thresholds based on class distribution and model learning status, improving pseudo-label quantity and quality. Experiments on private and public OCT datasets demonstrate that our method outperforms state-of-the-art methods.

Förderung

  • Key R&D Program of Shandong Province
  • National Natural Science Foundation of China
  • Key Technology Research and Development Program of Shandong Province
  • Science and Technology Innovation Program for Distinguished Young Scholars of Shandong Province Higher Education Institutions
  • Natural Science Foundation of Shandong Province
  • Shandong Provincial Natural Science Foundation for Distinguished Young Scholars
  • Taishan Scholar Project of Shandong Province
  • Special Funds for Distinguished Professors of Shandong Jianzhu University

Zitieren

APA

Meng, L., Xi, X., Han, J., Qiao, L., Yin, Y., & Chen, X. (2026). Difficulty-Aware Pseudo-Label Correction Network for Fine-Grained Classification of Choroidal Neovascularization in OCT Images. IEEE Transactions on Multimedia, 28, 7163-7175. https://doi.org/10.1109/tmm.2026.3673510

Vancouver

Meng L, Xi X, Han J, Qiao L, Yin Y, Chen X. Difficulty-Aware Pseudo-Label Correction Network for Fine-Grained Classification of Choroidal Neovascularization in OCT Images. IEEE Transactions on Multimedia. 2026;28:7163-7175. doi:10.1109/tmm.2026.3673510

BibTeX
@article{Meng2026,
  author = {Meng, Lingzhao and Xi, Xiaoming and Han, Jia and Qiao, Lishan and Yin, Yilong and Chen, Xinjian},
  title = {Difficulty-Aware Pseudo-Label Correction Network for Fine-Grained Classification of Choroidal Neovascularization in OCT Images},
  journal = {IEEE Transactions on Multimedia},
  year = {2026},
  volume = {28},
  pages = {7163--7175},
  doi = {10.1109/tmm.2026.3673510}
}
Kostenlos starten →