Skip to main content

Table 1 ROC scores in different datasets

From: A neural computational model for bottom-up attention with invariant and overcomplete representation

   Basic number ROC    Basic number ROC
   (feature maps)     (feature maps)  
   100(16) 0.5722    196(25) 0.7104
  our 196(25) 0.6427   our   
Bruce model 392(50) 0.6830 Weizmann model 256(36) 0.7282
data   576(64) 0.6864 data    
  Sal Tool 144(30) 0.6717   Sal Tool 144(30) 0.6233
  ful net 392(1) 0.5913   ful net 256(1) 0.4947
  rand net 392(50) 0.5965   rand net 256(36) 0.5011