top of page
Research & Projects
Deep Single Image Manipulation
GAN-based method for image manipulation tasks using a single image dataset
Unpaired Learning for
High Dynamic Range
Image Tone Mapping
Yael Vinker, Inbar Huberman, Raanan Fattal
-
Unsupervised GAN based approach for HDR tone mapping
-
The method produce artifact-free LDR renditions of highly challenging HDR images
![teaser (5).png](https://static.wixstatic.com/media/bc296b_cc8d9d20e6824221b4b91dc2abe5d683~mv2.png/v1/crop/x_718,y_3,w_667,h_355/fill/w_444,h_239,al_c,q_85,usm_0.66_1.00_0.01,enc_auto/teaser%20(5).png)
End-to-end single image
generators without GANs
Using non-adversarial methods to train generative models from a single image
MyWay- Smart Signs
T-Hack, multidisciplinary hackathon of the Technion, “smart cities” challenge
We created a smart road sign prototype which connects to the driver's navigation system and shows him only the correct driving direction, using light polarization (LCD) technology​
bottom of page