Keratoconus Screening Based on Deep Learning Approach of Corneal Topography.
Keratoconus Screening Based on Deep Learning Approach of Corneal Topography.
复制标题
DOI：
10.1167/tvst.9.2.53
复制
发表时间：
2020-09
影响因子：
3
通讯作者：
Wang IJ
中科院分区：
文献类型：
--
作者：
Kuo BI;Chang WY;Liao TS;Liu FY;Liu HY;Chu HS;Chen WL;Hu FR;Yen JY;Wang IJ
开发并比较基于角膜地形图的深度学习圆锥角膜检测算法，并用可视化方法进行验证。我们回顾性收集临床表现为圆锥角膜的研究组和有规律散光的对照组的角膜地形图。所有图像被分为训练和测试数据集。我们采用了三种卷积神经网络（CNN）模型进行学习。利用测试数据集对三种模型的性能进行了分析。此外，为了更好的区分和理解，我们展示了屈光度图像的逐像素区分特征和类别区分热图，用于可视化。共收集圆锥角膜170例，亚临床圆锥角膜28例，正常地形图156张。训练和测试数据集的准确率和损失的收敛表明，三种CNN模型都没有过拟合。所有CNN模型的灵敏度和特异度均在0.90以上，ResNet152模型的受试者工作特征曲线下面积达到0.995。VGG16模型中逐像素的判别特征和预测层的热图均显示其关注地形图梯度差最大的区域，这与眼科医生的诊断线索相对应。该模型对亚临床圆锥角膜有良好的预测作用，并与地形指数相关。DL模型对基于角膜地形图的圆锥角膜筛查具有较好的准确性。本研究中提到的可视化表明，该模型专注于适当的诊断区域，使深度学习的临床可解释性更容易被接受。这些高精度的CNN模型可以帮助眼科医生使用彩色编码的角膜地形图进行圆锥角膜筛查。
To develop and compare deep learning (DL) algorithms to detect keratoconus on the basis of corneal topography and validate with visualization methods. We retrospectively collected corneal topographies of the study group with clinically manifested keratoconus and the control group with regular astigmatism. All images were divided into training and test datasets. We adopted three convolutional neural network (CNN) models for learning. The test dataset was applied to analyze the performance of the three models. In addition, for better discrimination and understanding, we displayed the pixel-wise discriminative features and class-discriminative heat map of diopter images for visualization. Overall, 170 keratoconus, 28 subclinical keratoconus and 156 normal topographic pictures were collected. The convergence of accuracy and loss for the training and test datasets after training revealed no overfitting in all three CNN models. The sensitivity and specificity of all CNN models were over 0.90, and the area under the receiver operating characteristic curve reached 0.995 in the ResNet152 model. The pixel-wise discriminative features and the heat map of the prediction layer in the VGG16 model both revealed it focused on the largest gradient difference of topographic maps, which was corresponding to the diagnostic clues of ophthalmologists. The subclinical keratoconus was positively predicted with our model and also correlated with topographic indexes. The DL models had fair accuracy for keratoconus screening based on corneal topographic images. The visualization mentioned in the current study revealed that the model focused on the appropriate region for diagnosis and rendered clinical explainability of deep learning more acceptable. These high accuracy CNN models can aid ophthalmologists in keratoconus screening with color-coded corneal topography maps.
登录
查看更多内容
影响因子：
2.4
作者：
Chatzis, Nico;Hafezi, Farhad
通讯作者：
Hafezi, Farhad
影响因子：
13.7
作者：
de Sanctis, Ugo;Loiacono, Carlotta;Grignolo, Federico M.
通讯作者：
Grignolo, Federico M.
影响因子：
13.7
作者：
Hwang ES;Perez-Straziota CE;Kim SW;Santhiago MR;Randleman JB
通讯作者：
Randleman JB
影响因子：
14.9
作者：
Hinton, Geoffrey;Deng, Li;Kingsbury, Brian
通讯作者：
Kingsbury, Brian
影响因子：
4.1
作者：
Rabinowitz, YS;Yang, HY;Elashoff, J
通讯作者：
Elashoff, J