Skip to main content

Table 2 Comparison of nuclear segmentation between different fusion models and a previously published method

From: Fusion of encoder-decoder deep networks improves delineation of multiple nuclear phenotypes

Input size Recall Precision F1-Score Standard Error
Fused-ENet 0.94 0.88 0.91 1.1124e-06
Fused 3-module encoder, 3-module decoder 0.79 0.88 0.82 1.6170e-06
Multi-reference Graphcut (MRG) 0.75 0.85 0.79 unknown
  1. MRG was shown to out-perform other methods that include random forest
  2. The italic items illustrate the best results