Fig 1.
Flowchart of the segmentation method.
Fig 2.
(a) No pre-processing. (b) Bilateral Filter.
Fig 3.
The process for defining the area of interest.
Fig 4.
(a) Result of the ROI definition stage. (b) Image after dilation.
Fig 5.
The used U-Net architecture.
Fig 6.
Font: Adapted from [22].
Fig 7.
On the left, the original image of the mask, and on the right, the result of the network.
Table 1.
The results of segmentation of the edges of the retinal layers.
Fig 8.
Images from the exam “AMD _1057”, in the left column are the results of our method, the dataset annotation is in the right column.
Fig 9.
Images from the exam “AMD _1090”, in the left column are the results of our method, the dataset annotation is in the right column.
Fig 10.
Examples of edge detection failure.
(a) Exam B-scan 46 “AMD_1053”. (b) Exam B-scan 32 “AMD_1081”.
Table 2.
Comparison with related works.