EUVIMEDEUVIMED entdecken
Crossref

Computer-Aided Tear Film Breakup Time Test based Dry Eye Diagnosis using Explainable AI

Aditi Haresh Vyas, Dr. Vidhi Khanduja, Dr. Ravish Kinkhabwala

International Journal of Drug Delivery Technology · 15.05.2026 · 16 (27s)

DOI: 10.25258/ijddt.16.27s.28 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 15.05.2026. Verfasst von Aditi Haresh Vyas und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Early identification of eye disorders is important for preventing long term complications and maintaining visual comfort. Dry Eye Disease (DED) is a common ocular condition that can significantly affect the quality of life if it is not diagnosed and managed at an early stage. Among the clinical examinations used for DED assessment, the Tear Breakup Time (TBUT) test is widely used to evaluate the stability of the tear film. Motivated by the need for reliable and automated diagnostic support, this study proposes a deep learning based approach for the classification of TBUT images. A dataset of TBUT test images was collected for this study. Since the dataset is of small amount, data augmentation techniques were applied to increase the diversity of training samples and improve model performance. Prior to model training, several preprocessing steps were performed to enhance the quality of the images. Four pretrained convolutional neural network models-EfficientNetV2S, MobileNetV3, EfficientNetV2M, and ResNetV2 were employed for the classification task. In addition, Explainable Artificial Intelligence(XAI) techniques, namely Gradient-weighted Class Activation Mapping (Grad-CAM) and Local Interpretable Model-agnostic Explanations (LIME), were used to visualize and interpret the decision making process of the models. The inclusion of explainability helps improve transparency and supports the potential use of such systems in clinical environments. Experimental results indicate that EfficientNetV2S achieved the best performance among the evaluated models with accuracy of 95%. Overall, the findings demonstrate the potential of deep learning combined with XAI techniques for assisting clinicians in the diagnosis of DED using TBUT images.

Zitieren

APA

Vyas, A. H., Khanduja, D. V., & Kinkhabwala, D. R. (2026). Computer-Aided Tear Film Breakup Time Test based Dry Eye Diagnosis using Explainable AI. International Journal of Drug Delivery Technology, 16(27s). https://doi.org/10.25258/ijddt.16.27s.28

Vancouver

Vyas AH, Khanduja DV, Kinkhabwala DR. Computer-Aided Tear Film Breakup Time Test based Dry Eye Diagnosis using Explainable AI. International Journal of Drug Delivery Technology. 2026;16(27s). doi:10.25258/ijddt.16.27s.28

BibTeX
@article{Vyas2026,
  author = {Vyas, Aditi Haresh and Khanduja, Dr. Vidhi and Kinkhabwala, Dr. Ravish},
  title = {Computer-Aided Tear Film Breakup Time Test based Dry Eye Diagnosis using Explainable AI},
  journal = {International Journal of Drug Delivery Technology},
  year = {2026},
  volume = {16},
  number = {27s},
  doi = {10.25258/ijddt.16.27s.28}
}
Kostenlos starten →