Automated diagnosis of keratitis from low-quality slit-lamp images using an improved generative adversarial network
5 affiliations
- 1School of Electronic Engineering, Xi’an University of Posts and Telecommunications, Xi’an, China
- 2Ningbo Key Laboratory of Medical Research on Blinding Eye Diseases, Ningbo Eye Institute, Ningbo Eye Hospital, Wenzhou Medical University, Ningbo, China
- 3School of Ophthalmology and Optometry and Eye Hospital, Wenzhou Medical University, Wenzhou, China
- 4School of Mathematics and Statistics, Xidian University, Xi’an, China
- 5School of Computer Science and Technology, Xi’an University of Posts and Telecommunications, Xi’an, China
Abstract
Keratitis is a major cause of visual impairment worldwide, with timely and accurate diagnosis being critical to preventing irreversible vision loss. Existing deep learning (DL) models rely heavily on high-quality slit-lamp (HQS) images, limiting their applicability in real-world clinical scenarios where low-quality slit-lamp (LQS) images are frequently encountered due to acquisition-related artifacts. Here, we propose an automated keratitis diagnosis framework for LQS images (AKDF-LQSI) that integrates image enhancement with lesion-aware classification. A random degradation algorithm is designed to simulate common clinical artifacts, enabling paired LQS–HQS training. The framework incorporates an improved generative adversarial network to restore diagnostic features and a context-aware classifier to enhance lesion representation. AKDF-LQSI demonstrates robust performance when evaluated on a multicenter dataset of 10,498 slit-lamp images, achieving an AUC exceeding 0.949. This approach presents a practical solution for keratitis diagnosis from degraded images, with strong potential to reduce diagnostic disparities in resource-limited environments.
Related
- A rare case of keratitis caused by Scedosporium aurantiacum: a case report and review highlight the need for early microbiological diagnosis and treatment
- Privacy-preserving collaborative transfer learning with generative adversarial networks (CTL-GAN) for classification of dry eye disease
- Generative Artificial Intelligence Models in Breast Cancer Diagnosis: A Systematic Review
- Generative AI for Synthetic Medical Image Generation in Oncology: Addressing Data Scarcity in AI-Driven Cancer Diagnosis
- KGRD: a knowledge-graph-augmented automated reasoning framework for diagnosis and counselling of paediatric rare genetic disorders
BibTeX
@article{jiang2026automated,
title = {{Automated diagnosis of keratitis from low-quality slit-lamp images using an improved generative adversarial network}},
author = {Jiewei Jiang and Siyuan Song and Qihai Yao and Kuan Li and Tongtong Lin and Yu Xin and Yiwei Cui and Ning Wang and Yifan Bai and Xiaoci Luo and Wei Chen and Mingmin Zhu and Lang He and Zhongwen Li},
journal = {NPJ Digital Medicine},
year = {2026},
month = jul,
doi = {10.1038/s41746-026-02952-4},
url = {https://doi.org/10.1038/s41746-026-02952-4},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Automated diagnosis of keratitis from low-quality slit-lamp images using an improved generative adversarial network
AU - Jiewei Jiang
AU - Siyuan Song
AU - Qihai Yao
AU - Kuan Li
AU - Tongtong Lin
AU - Yu Xin
AU - Yiwei Cui
AU - Ning Wang
AU - Yifan Bai
AU - Xiaoci Luo
AU - Wei Chen
AU - Mingmin Zhu
AU - Lang He
AU - Zhongwen Li
PY - 2026
DA - 2026/07/03
JO - NPJ Digital Medicine
DO - 10.1038/s41746-026-02952-4
UR - https://doi.org/10.1038/s41746-026-02952-4
AB - Keratitis is a major cause of visual impairment worldwide, with timely and accurate diagnosis being critical to preventing irreversible vision loss. Existing deep learning (DL) models rely heavily on high-quality slit-lamp (HQS) images, limiting their applicability in real-world clinical scenarios where low-quality slit-lamp (LQS) images are frequently encountered due to acquisition-related artifacts. Here, we propose an automated keratitis diagnosis framework for LQS images (AKDF-LQSI) that integrates image enhancement with lesion-aware classification. A random degradation algorithm is designed to simulate common clinical artifacts, enabling paired LQS–HQS training. The framework incorporates an improved generative adversarial network to restore diagnostic features and a context-aware classifier to enhance lesion representation. AKDF-LQSI demonstrates robust performance when evaluated on a multicenter dataset of 10,498 slit-lamp images, achieving an AUC exceeding 0.949. This approach presents a practical solution for keratitis diagnosis from degraded images, with strong potential to reduce diagnostic disparities in resource-limited environments.
SN - 2398-6352
ER -