Related Experiment Video
Updated: May 1, 2026
10:25
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
Published on: November 11, 2022
11.7K
TearNET: Validation of a convolutional neural network for grading of tear ferning patterns using deep learning
Anantha Krishnan1, Pranay Gundeti1, Nagaraju Konda1
1School of Medical Sciences, Science Complex, University of Hyderabad, C.R Rao Road, Telangana, 500046 Gachibowli, Hyderabad, India.
Summary
A new deep learning algorithm, TearNET, accurately classifies tear ferning patterns, offering a promising tool for automated dry eye disease screening. This AI model shows potential for consistent and objective grading in clinical settings.
Area of Science:
- Ophthalmology
- Biomedical Engineering
- Artificial Intelligence
Background:
- Tear ferning patterns, microscopic crystallizations in tears, are influenced by tear biomolecular properties.
- These patterns can aid in screening for dry eye disease (DED).
- Current manual grading of tear ferning is subjective and inconsistent, necessitating objective methods.
Purpose of the Study:
- To validate TearNET, a convolutional neural network-based deep learning algorithm, for automated grading of tear ferning patterns.
- To assess TearNET's performance in classifying tear ferning patterns according to Rolando's grading system.
Main Methods:
- Tear samples from 80 healthy participants were collected and imaged.
- The TearNET model was trained on 70% of the samples and tested on 30%.
- Model performance was evaluated using sensitivity, specificity, recall, and F-scores.
Main Results:
- Tear ferning patterns showed significant variations with age and gender (p < 0.001).
- Room temperature and humidity did not significantly affect tear ferning patterns (p > 0.05).
- The TearNET model achieved 81% accuracy, with strong performance in classifying Types 3 and 4 ferning patterns.
Conclusions:
- The TearNET model shows significant promise for the automated classification of tear ferning patterns.
- The algorithm demonstrated effective training convergence and potential for clinical application in DED screening.
- TearNET offers a more objective and consistent approach to grading tear ferning compared to manual methods.