Related Experiment Video
Updated: Nov 11, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.1K
A method for the automatic detection of myopia in Optos fundus images based on deep learning
Zhengjin Shi1, Tianyu Wang1, Zheng Huang2,3,4
1School of Automation and Electrical Engineering, Shenyang Ligong University, Shenyang, China.
Summary
This study introduces a novel Myopia Detection Network (MDNet) for accurate myopia detection in fundus images. The proposed method achieves a Mean Absolute Error of 1.1150 D, demonstrating its effectiveness in diagnosing myopia.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Myopia detection is crucial for preventing vision loss and diagnosing related conditions like myopic retinopathy.
- Current detection methods may lack efficiency and accuracy, necessitating advanced automated solutions.
Purpose of the Study:
- To develop and validate a novel automated system, the Myopia Detection Network (MDNet), for enhanced myopia detection in Optos fundus images.
- To improve the accuracy and efficiency of myopia detection using a deep learning approach.
Main Methods:
- An automatic optic disc recognition method was employed for Region of Interest extraction and noise reduction.
- Data augmentation techniques were utilized to expand the dataset and mitigate overfitting.
- A Myopia Detection Network (MDNet) incorporating Attention Dense blocks was constructed for myopia detection.
Main Results:
- The MDNet achieved a Mean Absolute Error of 1.1150 D for Spherical Equivalent detection.
- The proposed method demonstrated feasibility and applicability for automated myopia detection in Optos fundus images.
Conclusions:
- The developed MDNet offers a promising automated solution for myopia detection.
- This approach can aid in early diagnosis and management of myopia and its complications.
