Table 8.

Cross-dataset analysis on the full retinal tissue segmentation results, presenting the mean (standard deviation) Dice coefficient (%) results. Three different training/testing combinations were included for each model, including: (i) The original using only the original dataset (O.D./O.D.), (ii) training adding the AMD dataset images (O.D. + AMD) and testing only on O.D. and (iii) training adding the AMD dataset images (O.D. + AMD) and testing only on AMD. O.D. denotes original healthy dataset imaged with the Spectralis device and the AMD dataset denotes the pathology dataset imaged with Bioptigen device.

Trained/Tested Data	Mask R-CNN	U-Net	FCN	DeeplabV3
O.D./O.D.	97.21 (0.19)	99.50 (0.01)	98.86 (0.01)	98.87 (0.03)
O.D. + AMD/O.D.	97.41 (0.26)	99.35 (0.01)	99.61 (0.01)	99.60 (0.01)
O.D. + AMD/AMD	96.40 (0.40)	99.55 (0.02)	99.83 (0.01)	99.83 (0.01)