Skip to main content

Table 2 Segmentation performance of different algorithms on the test set

From: Automated segmentation of choroidal neovascularization on optical coherence tomography angiography images of neovascular age-related macular degeneration patients based on deep learning

Method

Saliency-based method

U-net

Ours

AUC

0.9188(0.9185–0.9191)

0.9259(0.9257–0.9261)

0.9476(0.9473–0.9479)

ACC

0.9794(0.9792–0.9796)

0.9795(0.9792–0.9798)

0.9891(0.9889–0.9893)

SPE

0.9845(0.9840–0.9850)

0.9949(0.9943–0.9955)

0.9950(0.9945–0.9955)

SEN

0.7180(0.7175–0.7185)

0.6860(0.6854–0.6866)

0.7271(0.7265–0.7277)

IOU

0.4692(0.4690–0.4694)

0.5601(0.5598–0.5604)

0.5867(0.5864–0.5870)

DICE

0.6293(0.6290–0.6296)

0.7144(0.7140–0.7148)

0.7299(0.7295–0.7303)