An Artificial-Intelligence- and Telemedicine-Based Screening Tool to Identify Glaucoma Suspects from Color Fundus Imaging

Figure 3

Overall schematic representation of the model building. The two types of input images (RGB and transformed) fed into five neural networks are shown. Each network accepts a rescaled image (shown below the name of the architecture) and produces two probabilities. The ten resulting probabilities are then concatenated to form a feature array of size 10 and fed as input to a logistic model tree which acts as the final classifier in this system.