Skip to main content

Table 4 The model segmentation results with and without Multi input, deeper network and post-processing

From: Joint optic disc and cup segmentation based on densely connected depthwise separable convolution deep network

Method DRISHTI-GS dataset REFUGE dataset
\(DC_{disc}\,(mean)\) \(DC_{cup}\,(mean)\) \(DC_{disc}\,(mean)\) \(DC_{cup}\,(mean)\)
Without image pyramid input 0.9743 0.9076 0.9589 0.8880
Without multi-DDSC blocks 0.9746 0.8939 0.9575 0.8837
DDSC-Net 0.9779 0.9123 0.9592 0.8894
DDSC-NET + post-processing 0.9780 0.9123 0.9601 0.8903
  1. Italics indicate the best results