Home

Awesome

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

<img src="https://github.com/GauravBh1010tt/Controllable-Domain-Translation/blob/main/figs/figs/m2f_sub.jpg" width="700">

Forward-backward interpolation

<img src="https://github.com/GauravBh1010tt/Controllable-Domain-Translation/blob/main/figs/figs/fb_int.jpg" width="700">

Multi-attribute controllable generation

<img src="https://github.com/GauravBh1010tt/Controllable-Domain-Translation/blob/main/figs/figs/multi.jpg" width="500">

Multi-domain style transfer

<img src="https://github.com/GauravBh1010tt/Controllable-Domain-Translation/blob/main/figs/figs/multi_dom.jpg" width="700">

References-guided image synthesis

<img src="https://github.com/GauravBh1010tt/Controllable-Domain-Translation/blob/main/figs/figs/ref_celeb.jpg" width="700">

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} }