qAp/sartorius_cell_instance_segmentation_kaggle

Issues

• Challenge inference pipeline

  #13 opened 3 years ago by qAp
  16

• Watershed energy transform (WTN)

  #8 opened 3 years ago by qAp
  11

• Direction transform (DN)

  #2 opened 3 years ago by qAp
  20

• Watershed End-to-end (WN = DN + WTN)

  #10 opened 3 years ago by qAp
  5

• Semantic segmentation (Unet)

  #7 opened 3 years ago by qAp
  17

• Post-processing

  #9 opened 3 years ago by qAp
  2

• Re-generate Kaggle Datasets of ground truths

  #11 opened 3 years ago by qAp
  0

• DN inference breaks

  #12 opened 3 years ago by qAp
  1

• Rotational augmentation transforms for uvec

  #6 opened 3 years ago by qAp
  1

• What training target for the semantic segmentation stage?

  #4 opened 3 years ago by qAp
  2

• Increase width of cell border to separate cells?

  #3 opened 3 years ago by qAp
  1

• Save cell area to train.csv, then generate cv folds

  #5 opened 3 years ago by qAp
  1

• Cannot convert to torchscript when model's forward function instantiates a nn.Module.

  #1 opened 3 years ago by qAp
  0