Deep learning for predicting refractive error from multiple photorefraction images

• Published in: Biomedical engineering online Vol. 21; no. 1; pp. 1 - 55
• Main Authors: Xu, Daoliang, Ding, Shangshang, Zheng, Tianli, Zhu, Xingshuai, Gu, Zhiheng, Ye, Bin, Fu, Weiwei
• Format: Journal Article
• Language: English
• Published: London BioMed Central Ltd 08.08.2022; BioMed Central; BMC
• Subjects: Accuracy; Algorithms; Artificial neural networks; Cellular telephones; Computer-aided medical diagnosis; Convolutional neural network; Deep learning; Equivalence; Error correction & detection; Error detection; Eye; Feature extraction; Image processing; Machine learning; Medical screening; Methods; Myopia; Neural networks; Performance evaluation; Photorefraction; Recurrent neural networks; Refractive error; Refractive errors; China
• ISSN: 1475-925X; 1475-925X
• DOI: 10.1186/s12938-022-01025-3

Refractive error detection is a significant factor in preventing the development of myopia. To improve the efficiency and accuracy of refractive error detection, a refractive error detection network (REDNet) is proposed that combines the advantages of a convolutional neural network (CNN) and a recurrent neural network (RNN). It not only extracts the features of each image, but also fully utilizes the sequential relationship between images. In this article, we develop a system to predict the spherical power, cylindrical power, and spherical equivalent in multiple eccentric photorefraction images. The results show that the mean absolute error (MAE) values of the spherical power, cylindrical power, and spherical equivalent can reach 0.1740 D (diopters), 0.0702 D, and 0.1835 D, respectively. This method demonstrates a much higher accuracy than those of current state-of-the-art deep-learning methods. Moreover, it is effective and practical.