Sun Ling-Chun, Pao Shu-I, Huang Ke-Hao, Wei Chih-Yuan, Lin Ke-Feng, Chen Ping-Nan
School of Medicine, National Defense Medical Center, Taipei, Taiwan.
Department of Ophthalmology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan.
Graefes Arch Clin Exp Ophthalmol. 2023 May;261(5):1399-1412. doi: 10.1007/s00417-022-05919-9. Epub 2022 Nov 28.
To determine whether a deep learning approach using generative adversarial networks (GANs) is beneficial for the classification of retinal conditions with Optical coherence tomography (OCT) images.
Our study utilized 84,452 retinal OCT images obtained from a publicly available dataset (Kermany Dataset). Employing GAN, synthetic OCT images are produced to balance classes of retinal disorders. A deep learning classification model is constructed using pretrained deep neural networks (DNNs), and outcomes are evaluated using 2082 images collected from patients who visited the Department of Ophthalmology and the Department of Endocrinology and Metabolism at the Tri-service General Hospital in Taipei from January 2017 to December 2021.
The highest classification accuracies accomplished by deep learning machines trained on the unbalanced dataset for its training set, validation set, fivefold cross validation (CV), Kermany test set, and TSGH test set were 97.73%, 96.51%, 97.14%, 99.59%, and 81.03%, respectively. The highest classification accuracies accomplished by deep learning machines trained on the synthesis-balanced dataset for its training set, validation set, fivefold CV, Kermany test set, and TSGH test set were 98.60%, 98.41%, 98.52%, 99.38%, and 84.92%, respectively. In comparing the highest accuracies, deep learning machines trained on the synthesis-balanced dataset outperformed deep learning machines trained on the unbalanced dataset for the training set, validation set, fivefold CV, and TSGH test set.
Overall, deep learning machines on a synthesis-balanced dataset demonstrated to be advantageous over deep learning machines trained on an unbalanced dataset for the classification of retinal conditions.
确定使用生成对抗网络（GAN）的深度学习方法是否有助于通过光学相干断层扫描（OCT）图像对视网膜疾病进行分类。
我们的研究使用了从公开可用数据集（克曼尼数据集）获得的84452张视网膜OCT图像。利用GAN生成合成OCT图像以平衡视网膜疾病的类别。使用预训练的深度神经网络（DNN）构建深度学习分类模型，并使用2017年1月至2021年12月期间在台北三军总医院眼科和内分泌代谢科就诊的患者收集的2082张图像评估结果。
在不平衡数据集上训练的深度学习机器在其训练集、验证集、五重交叉验证（CV）、克曼尼测试集和三军总医院测试集上实现的最高分类准确率分别为97.73%、96.51%、97.14%、99.59%和81.03%。在合成平衡数据集上训练的深度学习机器在其训练集、验证集、五重CV、克曼尼测试集和三军总医院测试集上实现的最高分类准确率分别为98.60%、98.41%、98.52%、99.38%和84.92%。在比较最高准确率时，在合成平衡数据集上训练的深度学习机器在训练集、验证集、五重CV和三军总医院测试集方面优于在不平衡数据集上训练的深度学习机器。
总体而言，对于视网膜疾病的分类，在合成平衡数据集上的深度学习机器比在不平衡数据集上训练的深度学习机器更具优势。
