Using AutoDIAL framework for training US-CT end to end network working on real US images. Based on the paper: Maria Carlucci, Fabio, et al. "AutoDIAL: Automatic DomaIn Alignment Layers." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017 Datasets should be stored in the following hierarchy: datasets--->Name_of_data_set--->source--->samples datasets--->Name_of_data_set--->source--->labels datasets--->Name_of_data_set--->target--->samples For training run from terminal: python main.py --phase=train --dataset_dir=Name_of_data_set (...any other parameter) For test run from terminal: python main.py --phase=test --dataset_dir=Name_of_data_set --domain=target For displaying results in tensorboard run from terminal: tensorboard --logdir=./logs
-
Notifications
You must be signed in to change notification settings - Fork 2
ortalsen/DomainTransfer
About
Using AutoDIAL framework for training US-CT end to end network working on real US images
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published