EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A novel feature-agnostic approach for glaucoma detection in fundus images

Nadia Rasool, Zulkaif Sajjad, Furqan Shaukat, Syeda Filzah Bukhari, Mahmood Ali, Farah Akhtar

Journal of the Pakistan Medical Association · 20.06.2026 · 1-18

DOI: 10.47391/jpma.31245 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of the Pakistan Medical Association, online erschienen am 20.06.2026. Verfasst von Nadia Rasool und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Glaucoma, Al-Shifa Trust Eye Hospital, Rawalpindi, Pakistan. Hauptthemen laut MEDLINE: Glaucoma, Fundus Oculi, Deep Learning, Image Interpretation, Computer-Assisted. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

To design and validate a fully automated, feature-agnostic deep learning approach for the early detection of glaucoma from fundus images.

Method

The retrospective study was conducted at Al-Shifa Trust Eye Hospital, Rawalpindi, Pakistan, using publicly available retinal fundus image datasets from October 2024 to January 2025, and comprised 1,707 fundus images from five publicly available retrospective datasets. A computer-aided detection system was employed based on state-of-the-art deep convolutional neural networks, including EfficientNetV2b0, Xception, InceptionV3, Visual Geometry Group and ResNet50. The images were labelled either as glaucomatous or healthy after they were pre-processed through cropping, normalisation and data augmentation. The models were fine-tuned using transfer learning, and evaluated using standard metrics, such as accuracy, precision, recall, F1-score and area under the curve, which were calculated using Python-based statistical libraries.

Results

Among the tested models, EfficientNetV2b0 achieved the best performance with an area under the curve of 0.98 and an accuracy of 93%. The best-performing model achieved a sensitivity/recall of 97%, precision of 89% and F1-score of 93%, indicating reliable performance for glaucoma classification. The robustness of the proposed method was validated across multiple datasets, ensuring its generalisability in diverse clinical scenarios.

Conclusion

The proposed deep learning-based approach provided a reliable and efficient method for early glaucoma detection. Its automation and high accuracy made it suitable for use in mass screening programmes and under-resourced clinical environments, potentially reducing the burden on ophthalmologists and enabling timely intervention. Key Words: Glaucoma detection, Fundus imaging, Computer-aided diagnosis, Optic nerve head, optical disc, Optical cup.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

GlaucomaFundus OculiDeep LearningImage Interpretation, Computer-AssistedHumansConvolutional Neural NetworksRetrospective StudiesDetection Algorithms

Stichwörter der Autor:innen

Glaucoma detection, Fundus imaging, Computer-aided diagnosis, Optic nerve head,…

Autor:innen und Institutionen

  • Department of Glaucoma, Al-Shifa Trust Eye Hospital, Rawalpindi, Pakistan.
  • Department of Electrical, University of Engineering and Technology, Taxila, Pakistan.

Zitieren

APA

Rasool, N., Sajjad, Z., Shaukat, F., Bukhari, S. F., Ali, M., & Akhtar, F. (2026). A novel feature-agnostic approach for glaucoma detection in fundus images. Journal of the Pakistan Medical Association, 1-18. https://doi.org/10.47391/jpma.31245

Vancouver

Rasool N, Sajjad Z, Shaukat F, Bukhari SF, Ali M, Akhtar F. A novel feature-agnostic approach for glaucoma detection in fundus images. Journal of the Pakistan Medical Association. 2026:1-18. doi:10.47391/jpma.31245

BibTeX
@article{Rasool2026,
  author = {Rasool, Nadia and Sajjad, Zulkaif and Shaukat, Furqan and Bukhari, Syeda Filzah and Ali, Mahmood and Akhtar, Farah},
  title = {A novel feature-agnostic approach for glaucoma detection in fundus images},
  journal = {Journal of the Pakistan Medical Association},
  year = {2026},
  pages = {1--18},
  doi = {10.47391/jpma.31245}
}
Kostenlos starten →