Segmentation of Laser Marks of Diabetic Retinopathy in the Fundus Photographs Using Lightweight U-Net

Table 1

Comparison of methods on the proposed dataset.

Architecture	Sensitivity	Specificity	Accuracy	DSC	AUC
Lightweight U-Net	93.55%	92.16%	92.25%	70.59%	0.9798
Lightweight U-Net with macular regions covered	93.55%	92.38%	92.46%	70.76%	0.9804
Light weight U-Net with optic discs covered	94.15%	92.58%	92.68%	71.01%	0.9818
Lightweight U-Net with both post-processing procedures	94.16%	92.82%	92.90%	71.18%	0.9824

TP: the number of pixels correctly classified as laser marks; †FN: the number of pixels mistakenly classified as background; ‡ TN: the number of pixels correctly classified as background; § FP: the number of pixels mistakenly classified as laser marks.