Deep Imbalanced Regression Model for Predicting Refractive Error from Retinal Photos.

ABSTRACT:

Purpose

Recent studies utilized ocular images and deep learning (DL) to predict refractive error and yielded notable results. However, most studies did not address biases from imbalanced datasets or conduct external validations. To address these gaps, this study aimed to integrate the deep imbalanced regression (DIR) technique into ResNet and Vision Transformer models to predict refractive error from retinal photographs.

Design

Retrospective study.

Subjects

We developed the DL models using up to 103 865 images from the Singapore Epidemiology of Eye Diseases Study and the United Kingdom Biobank, with internal testing on up to 8067 images. External testing was conducted on 7043 images from the Singapore Prospective Study and 5539 images from the Beijing Eye Study. Retin

SUBMITTER: Yew SME 

PROVIDER: S-EPMC11808727 | biostudies-literature | 2025 Mar-Apr

REPOSITORIES: biostudies-literature