Controllable-Domain-Translation
Code for our paper Learning Style Subspaces for Controllable Unpaired Domain Translation
Train
$python main.py --num_domains=2 --mode='train' --real_img_dir='data/celeb_hq/train'
Inference
$python main.py --num_domains=2 --mode='sample' --real_img_dir='data/celeb_hq/train'
Results of Controllable domain translation
Forward-backward interpolation
Multi-attribute controllable generation
Multi-domain style transfer
References-guided image synthesis
BibTex:-
@InProceedings{Bhatt_2023_WACV, author = {Bhatt, Gaurav and Balasubramanian, Vineeth N.}, title = {Learning Style Subspaces for Controllable Unpaired Domain Translation}, booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)}, month = {January}, year = {2023}, pages = {4220-4229} }