Myopia prediction for children and adolescents via time-aware deep learning

4Citations
Citations of this article
11Readers
Mendeley users who have this article in their library.

This article is free to access.

Abstract

This is a retrospective analysis. Quantitative prediction of the children’s and adolescents’ spherical equivalent based on their variable-length historical vision records. From October 2019 to March 2022, we examined uncorrected visual acuity, sphere, astigmatism, axis, corneal curvature and axial length of 75,172 eyes from 37,586 children and adolescents aged 6–20 years in Chengdu, China. 80% samples consist of the training set, the 10% form the validation set and the remaining 10% form the testing set. Time-Aware Long Short-Term Memory was used to quantitatively predict the children’s and adolescents’ spherical equivalent within two and a half years. The mean absolute prediction error on the testing set was 0.103 ± 0.140 (D) for spherical equivalent, ranging from 0.040 ± 0.050 (D) to 0.187 ± 0.168 (D) if we consider different lengths of historical records and different prediction durations. Time-Aware Long Short-Term Memory was applied to captured the temporal features in irregularly sampled time series, which is more in line with the characteristics of real data and thus has higher applicability, and helps to identify the progression of myopia earlier. The overall error 0.103 (D) is much smaller than the criterion for clinically acceptable prediction, say 0.75 (D).

Cite

CITATION STYLE

APA

Huang, J., Ma, W., Li, R., Zhao, N., & Zhou, T. (2023). Myopia prediction for children and adolescents via time-aware deep learning. Scientific Reports, 13(1). https://doi.org/10.1038/s41598-023-32367-0

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free