Related Experiment Video
Updated: Aug 2, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.8K
Development of a robust eye exam diagnosis platform with a deep learning model
1Department of Information and Telecommunication Engineering, Gangeung-Wonju National University, Wonju, Korea.
Summary
This study introduces a smartphone-based eye exam using ResNet-18 for accurate eye disease detection. The developed model achieved high accuracy, offering a stable and predictable diagnostic tool.
Area of Science:
- Ophthalmology
- Computer Vision
- Medical Imaging
Background:
- Traditional eye exams rely on expensive and unreliable equipment.
- Early detection of eye diseases is crucial for effective treatment.
Purpose of the Study:
- To develop a stable and predictable model for automatic eye disease identification using smartphone-based optometric lenses.
- To create a new method for accurate and consistent eye disease diagnosis.
Main Methods:
- Utilized ResNet-18 models pre-trained on ImageNet data.
- Employed a dataset representing the target group domain for training and validation.
- Separated training and testing datasets to validate model performance.
Main Results:
- The proposed model achieved high training accuracy of 99.1%.
- Validation accuracy reached 96.9%, indicating robust performance.
- Demonstrated stable and predictable eye disease discrimination.
Conclusions:
- The developed model provides a robust and stable method for eye disease discrimination.
- Smartphone-based eye examination offers a promising alternative to traditional methods.
