Per-Lesion Mask-Supervised Attention for Multi-Label Diabetic Retinopathy Lesion Detection
Diagnostics · 05.11.2026 · 16 (21) · 3662
DOI: 10.3390/diagnostics16213662 ↗Kurz gefasst
Fachartikel in der Zeitschrift Diagnostics, online erschienen am 05.11.2026. Verfasst von Wael Alghamdi und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah 21589, Saudi Arabia. Die Arbeit nennt 99 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: King Abdulaziz University.
Abstract (Originalsprache)
Background/Objectives: Deep learning models for diabetic retinopathy (DR) usually report a severity grade, and their attention maps explain that grade rather than individual lesions. We aimed to detect microaneurysms, hemorrhages, hard exudates and soft exudates (cotton-wool spots) separately, each with an attention map trained to fall on that lesion.
Methods
An EfficientNet-B4 and Transformer encoder, pretrained on severity grades, feeds four lesion queries that each return a score and an attention map; a localization loss trains each map toward its lesion mask. We trained it on DDR and TJDR, tested it on their test sets and on two external sets (IDRiD and e-ophtha), and compared it with a U-Net.
Results
The mean areas under the curve (AUCs) were 0.93–0.94 for hemorrhages, 0.90–0.96 for hard exudates and 0.81–0.90 for soft exudates; microaneurysms reached 0.91–0.96, except on DDR (0.58). We analyze this drop in detail. Attention ranked lesion pixels above the rest of the retina with an AUC-Judd of 0.948–0.996. Without the localization loss, detection was slightly lower, and attention moved to regions outside the lesions that still raised the prediction. The model detected lesions better than U-Net (mean AUC of 0.876 against 0.853) and, unlike the latter, also learned from images that have labels but no masks: with masks on half of the training images, detection and AUC-Judd were nearly unchanged.
Conclusions
Supervising each lesion’s attention with its own mask gives a per-lesion score and a map that falls on that lesion, without reducing detection.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Computer Science, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah 21589, Saudi Arabia.
- Department of Information Systems, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah 21589, Saudi Arabia.
ORCID-Profile: Wael Alghamdi, Fawaz Alsolami, Mostafa Saleh
Förderung
- King Abdulaziz University
Zitieren
APA
Alghamdi, W., Alsolami, F., & Saleh, M. (2026). Per-Lesion Mask-Supervised Attention for Multi-Label Diabetic Retinopathy Lesion Detection. Diagnostics, 16(21), 3662. https://doi.org/10.3390/diagnostics16213662
Vancouver
Alghamdi W, Alsolami F, Saleh M. Per-Lesion Mask-Supervised Attention for Multi-Label Diabetic Retinopathy Lesion Detection. Diagnostics. 2026;16(21):3662. doi:10.3390/diagnostics16213662
BibTeX
@article{Alghamdi2026,
author = {Alghamdi, Wael and Alsolami, Fawaz and Saleh, Mostafa},
title = {Per-Lesion Mask-Supervised Attention for Multi-Label Diabetic Retinopathy Lesion Detection},
journal = {Diagnostics},
year = {2026},
volume = {16},
number = {21},
pages = {3662},
doi = {10.3390/diagnostics16213662}
}
EUVIMED