Weakly supervised underwater fish segmentation using affinity lcfcn

HIGHLIGHTS

• who: Issam H. Laradji from the Canada James Cook University have published the paper: Weakly supervised underwater fish segmentation using affinity LCFCN, in the Journal: Scientific Reports Scientific Reports
• what: The authors propose A-LCFCN, which extends a fully convolutional neural_network with an affinity-based module that is trained using the LCFCN loss. The aim of the training strategy is to learn to output a single blob per fish in the image using point-level annotations (Fig 1). The authors evaluate the models on two splits of the DeepFish d ­ ataset32, FishSeg and FishLoc to . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or Email

  Password

  Remember me

  Lost your password?

  If you don't have an account, you can create one here.