Wonder

Generative Adversarial Networks

Delivered October 31, 2019. Contributor: Joseph M.

Goals

To find information on the use of Generative Adversarial Networks in changing facial attributes such as closing eyes and changing facial expressions on demand in a photo-realistic way. 

Early Findings

Generative Adversarial NetworksAccording to Tech Target, Generative Adversarial Networks (GAN) is a "machine learning model in which two neural networks compete with each other to become more accurate in their predictions."
GAN is used in a variety of ways that include photograph editing, face aging, photo blending, among many other uses. 
There are numerous repos and peer reviewed research papers on using GAN to manipulate facial attributes. 
ReposExamples of repos on the use of GAN in the manipulation of facial attributes include:
 SpatialAttentionGAN.
 Gan-toolkit. 
 Keras-GAN. 
Research PapersExamples of research papers on the use of GAN in the manipulation of facial attributes include:
AttGAN: Facial Attribute Editing by Only Changing What You Want.
Mask-aware Photorealistic Face Attribute Manipulation.
 Semi-Latent GAN: Learning to generate and modify facial images from attributes.

Proposed next steps:

You need to be the project owner to select a next step.