Nav apraksta

Hao SUN 7d73cf04f0 migrate to new scipy 4 gadi atpakaļ
datasets aa2d50d2df added data 6 gadi atpakaļ
imgs 9d0cb59da7 initial commit 6 gadi atpakaļ
.gitattributes c0c20cf92b Initial commit 6 gadi atpakaļ
.gitignore aa2d50d2df added data 6 gadi atpakaļ
README.md 7d73cf04f0 migrate to new scipy 4 gadi atpakaļ
dice_score.m 9d0cb59da7 initial commit 6 gadi atpakaļ
dice_score_multiple.m 9d0cb59da7 initial commit 6 gadi atpakaļ
main.py 7d73cf04f0 migrate to new scipy 4 gadi atpakaļ
model.py 7d73cf04f0 migrate to new scipy 4 gadi atpakaļ
module.py 9d0cb59da7 initial commit 6 gadi atpakaļ
ops.py 9d0cb59da7 initial commit 6 gadi atpakaļ
prepareimgs.m 9d0cb59da7 initial commit 6 gadi atpakaļ
requirements.txt 9d0cb59da7 initial commit 6 gadi atpakaļ
utils.py 7d73cf04f0 migrate to new scipy 4 gadi atpakaļ

README.md

CycleGAN

As proposed by Jun-Yan Zhu in Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks.

Prerequisites

  • tensorflow r1.1
  • numpy 1.11.0
  • scipy 0.17.0
  • pillow 3.3.0

Train

python main.py --dataset_dir=med-image

Test

python main.py --dataset_dir=med-image --phase=test --which_direction=AtoB

References