Deep learning for objective OCTA detection of diabetic retinopathy

David Le; Minhaj Nur Alam; Jennifer I. Lim; R. V. P. Chan; Xincheng Yao

doi:10.1117/12.2546586

Translator Disclaimer

19 February 2020 Deep learning for objective OCTA detection of diabetic retinopathy

David Le, Minhaj Nur Alam, Jennifer I. Lim, R. V. P. Chan, Xincheng Yao

Author Affiliations +

Proceedings Volume 11218, Ophthalmic Technologies XXX; 112181P (2020) https://doi.org/10.1117/12.2546586
Event: SPIE BiOS, 2020, San Francisco, California, United States

Abstract

Early detection of diabetic retinopathy (DR) is an essential step to prevent vision losses. This study is the first effort to explore convolutional neural networks (CNNs) for transfer-learning based optical coherence tomography angiography (OCTA) detection and classification of DR. We employed transfer-learning using a pre-trained CNN, VGG16, based on the ImageNet dataset for classification of OCTA images. To prevent overfitting, data augmentation, e.g. rotations, flips, and zooming, and 5-fold cross-validation were implemented. A dataset comprising of 131 OCTA images from 20 control, 17 diabetic patients without DR (NoDR), and 60 nonproliferative DR (NPDR) patients were used for preliminary validation. Best classification performance was achieved with fine-tuning nine layers of the sixteen-layer CNN model.

Citation Download Citation

David Le, Minhaj Nur Alam, Jennifer I. Lim, R. V. P. Chan, and Xincheng Yao "Deep learning for objective OCTA detection of diabetic retinopathy", Proc. SPIE 11218, Ophthalmic Technologies XXX, 112181P (19 February 2020); https://doi.org/10.1117/12.2546586

ACCESS THE FULL ARTICLE