Figure 5 | Scientific Reports

Figure 5

From: Deep learning models for screening of high myopia using optical coherence tomography

Figure 5

Visual explanations generated by Grad-CAM +  + on OCT scans of myopia class. The Grad-CAM +  + image show that the deep learning model identified some of the characteristic features of high myopia. The deep learning model accurately identified severe curvature of high myopic eye for all of the OCT images. The model also identified vascular microfolds (b), peripapillary artrophy (c, d and e), chorioretinal macular atrophy (e), Retinal cysts and paravascular lamellar hole (g) and retinoschisis (g and h).

Back to article page