Figure 5
From: Automatic choroidal segmentation in OCT images using supervised deep learning methods
Example segmentations for the standard U-Net architecture (no contrast enhancement). Solid lines correspond to the true boundary positions and the dotted lines represent the predicted locations. Left: three cases of successful delineation of all boundaries with close agreement between the predictions and truths. Right: three cases of failure on the CSI boundary with observable differences between the prediction and truth.
