notes.md

Notes

links http://stacom.cardiacatlas.org/ https://acdc.creatis.insa-lyon.fr/# http://www.cardiacatlas.org/challenges/

(ED-ES info)[https://www.medicalnewstoday.com/articles/325498.php]

Todo

1 - Add callbacks for training, start with end of epoch (so we can run eval) 2 - Data augmentation

Data processig

Initial steps:

1 - convert nib to numpy array (Done)

Build dataloader:

2 - need to deal with different images sizes (done) 3 - build data loader with image normalising (done)

Later: 4 - build in data augmentation

Model

1 - build model and archetecture and or find one on the interweb

Build Training (Done)

2 - save checkpoints docs

Build Eval

Tensorboard

pytorch tensorboard

1 - Add metrics to tensorboard (Done - training) 2 - Add predicions (images) to tensorboard