Skip to main content
Fig. 3 | BMC Medical Imaging

Fig. 3

From: Use of multimodal dataset in AI for detecting glaucoma based on fundus photographs assessed with OCT: focus group study on high prevalence of myopia

Fig. 3

Inception v3, Inception ResNet v2, and Xception model comparison. Three CNN models: the Xception, Inception v3, and Inception ResNet v2 were compared. All of the models can achieve a 94% micro-average AUROC with tenfold cross-validation. Only the Xception model confusion matrix is shown because the model is the latest version of the three. Although the overall accuracy can reach 87.09%, the sensitivity for the PPG group was low. Only 33 of 109 images were correctly predicted as being PPG. CNN, convolutional neural network; AUROC, area under the receiver operating characteristic curve; PPG, pre-perimetric glaucoma

Back to article page

BMC Medical Imaging

ISSN: 1471-2342

Contact us