Skip to main content
. 2019 Nov 11;10(12):6204–6226. doi: 10.1364/BOE.10.006204

Fig. 9.

Fig. 9.

Predictions of the best model compared to ophthalmologist-based truth. Examples of correct, incorrect, and borderline predictions included the ophthalmologist truth and the best model prediction probabilities. The prediction probability value was between 0 and 1, and the category of an image identified by the algorithm was determined by the largest value. CNV: choroidal neovascularization; DME: diabetic macular edema.