Skip to main content
. 2019 Jan 25;10(2):892–913. doi: 10.1364/BOE.10.000892

Fig. 8.

Fig. 8

Example of the classification of the VGG19_TL network for images from the test set of the ESPERANZA database. (a, b) True negative examples, the human evaluators and the algorithm identified the images as normal. (c, d) True positives examples, the human evaluators and the algorithm identified the images as glaucoma. (e, f) False negatives examples. The human evaluation identified both images as glaucoma but the algorithm labeled them as normal. (g, h) False positives examples. The human evaluators marked the images as normal but the algorithm classified them as glaucoma.