Golden
Pix2pixHD

Pix2pixHD

A method for high-resolution photorealistic image-to-image translation

All edits

Edits on 14 Mar 2019
Jed Christiansen
Jed Christiansen edited on 14 Mar 2019 10:51 pm
Edits made to:
Infobox (+1 properties)

Infobox

Edits on 9 Jan 2019
Daniel Frumkin"Wrote article, added resources, categories, and related topics. "
Daniel Frumkin edited on 9 Jan 2019 3:33 pm
Edits made to:
Article (+2 images) (+1444/-9 characters)
Further reading (+2 rows) (+8 cells) (+311 characters)
Documentaries, videos and podcasts (+1 rows) (+2 cells) (+62 characters)
Companies (+1 rows) (+4 cells) (+81 characters)
Categories (+2 topics)
Related Topics (+4 topics)

Article

PIx2pixHDPix2pixHD is a method for synthesizing high resolution photo-realistic images from semantic label maps using conditional generative adversarial networks (CGANs). It can generate high resolution image results with a novel adversarial loss, as well as new multi-scale generator and discriminator architectures.



In less technical terms, pix2pixHD is a straightforward way to generate high-resolution images with nearly endless options to change small and large details about the images. This is done by drawing on a label map (Fig. 1) and then translating the drawings using GANs to produce HD image outputs (Fig. 2).



In Fig. 2, you can see examples where significant changes were made. On the left, some of the cars have different colors, the shadow has been removed from the sidewalk, and the ground has been changed from asphalt to bricks. On the right, trees have been added across the top of the image and the sidewalk on the right is lighter with a green tint. There are more small details changed in each image that you can notice upon closer inspection.



The pix2pixHD methodology was originally introduced in 2017 in an academic paper by Ph.D. researchers from the University of California, Berkeley in coordination with NVIDIA Corporation. A revised version of the paper was published in August 2018. The code is available on github.



A paper titled Everybody Dance Now went on to modify the adversarial training setup of pix2pixHD in order to produce temporally coherent video frames such that the moves of a dancer in a source video were translated onto a target who appears to be doing the same dance moves in a second video but it is in fact a generated video. 



Companies

Company
CEO
Location
Products/Services

NVIDIA Corporation

Jensen Huang

Santa Clara, US

Computing, AI, and computer graphics

Further reading

Title
Author
Link
Type

Computers can make you dance, see how "Everybody can dance now!"

Samhita Alla

Web

Everybody Dance Now

Caroline Chan, Shiry Ginosar, Tinghui Zhou, Alexei A. Efros

Academic paper

Documentaries, videos and podcasts

Title
Date
Link

Categories

Related Topics

Edits on 7 Aug 2018
Golden AI"Linkify text links in standard tables"
Golden AI edited on 7 Aug 2018 11:18 pm
Edits made to:
Further reading (+36/-36 characters)
Documentaries, videos and podcasts (+60/-60 characters)

Further reading

Author
Title
Link
Type

Ting-Chun Wang, Ming-Yu Liu, Jun-Yan Zhu, Andrew Tao, Jan Kautz and Bryan Catanzaro

High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

Academic paper

Documentaries, videos and podcasts

Title
Date
Link

High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

30 November 2017

Edits on 5 Jun 2018
Golden AI"Corrections"
Golden AI edited on 5 Jun 2018 11:29 pm
Edits made to:
Further reading (+5/-5 characters)

Further reading

Author
Title
Link
Type

Ting-Chun Wang, Ming-Yu Liu, Jun-Yan Zhu, Andrew Tao, Jan Kautz and Bryan Catanzaro

High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

Academic Paperpaper

Edits on 1 Jun 2018
Golden AI"Merging standard tables"
Golden AI edited on 1 Jun 2018 3:35 am
Edits made to:
Academic papers (-1 rows) (-3 cells) (-198 characters)
Further reading (+1 rows) (+4 cells) (+212 characters)

Academic papers

Author
Title
Link

Ting-Chun Wang, Ming-Yu Liu, Jun-Yan Zhu, Andrew Tao, Jan Kautz and Bryan Catanzaro

High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

Further reading

Author
Title
Link
Type

Ting-Chun Wang, Ming-Yu Liu, Jun-Yan Zhu, Andrew Tao, Jan Kautz and Bryan Catanzaro

High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

Academic Paper

Edits on 21 Jan 2018
Melanie Manipula"added thumbnail"
Melanie Manipula edited on 21 Jan 2018 8:21 pm
Edits made to:
Topic thumbnail

Pix2pixHD

A method for high-resolution photorealistic image-to-image translation

Edits on 21 Jan 2018
Melanie Manipula
Melanie Manipula edited on 21 Jan 2018 2:51 am
Edits made to:
Description (+70 characters)
Article (+313 characters)
Academic papers (+1 rows)
Documentaries, videos and podcasts (+1 rows)
Topic thumbnail

Pix2pixHD

A method for high-resolution photorealistic image-to-image translation

Article



PIx2pixHD is a method for synthesizing high resolution photo-realistic images from semantic label maps using conditional generative adversarial networks (CGANs). It can generate high resolution image results with a novel adversarial loss, as well as new multi-scale generator and discriminator architectures.

Academic papers

Author
Title
Link

Ting-Chun Wang, Ming-Yu Liu, Jun-Yan Zhu, Andrew Tao, Jan Kautz and Bryan Catanzaro

High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

Documentaries, videos and podcasts

Title
Date
Link

High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs

30 November 2017

Melanie Manipula"Initial topic creation"
Melanie Manipula created this topic on 21 Jan 2018 1:35 am
Edits made to:
Topic thumbnail

 Pix2pixHD

A method for high-resolution photorealistic image-to-image translation

Article



No more activity to show.