Figure 2 | Scientific Reports

Figure 2

From: Cross-instrument optical coherence tomography-angiography (OCTA)-based prediction of age-related macular degeneration (AMD) disease activity using artificial intelligence

Figure 2

The network structure of the classifier used in our experiments. We use an ImageNet pre-trained EfficientNet-b5 as the backbone with a few modifications. An additional convolution layer is inserted to handle the four-channel inputs of OCTA and the output channel of the final fully connected layer’s output channels are reduced to four, aligning with the number of our desired categories.

Back to article page