"No one is harder on a talented person than the person themselves" - Linda Wilkinson ; "Trust your guts and don't follow the herd" ; "Validate direction not destination" ;

April 03, 2021

GAN - Aging Papers

AgingMapGAN (AMGAN): High-Resolution Controllable Face Aging with Spatially-Aware Conditional GANs

Key Notes

  • A model capable of individually transforming the local aging signs. Aging process of the different parts of the face.
  • Patch-based approach to enable inference on high-resolution images while keeping the computational cost of training the model low


Dataset - Flicker Faces High-Quality Dataset (FFHQ) 

More read - Link

PFA-GAN: Progressive Face Aging with Generative Adversarial Network

Key Notes

  • Progressive face aging framework based on generative adversarial network (PFA-GAN) to model the face age progression in a progressive way
  • We introduce an age estimation loss to take into account the age distribution for an improved aging accuracy


Triple-GAN: Progressive Face Aging with Triple Translation Loss

Key Notes

  • GAN adopts triple translation loss to model the strong interrelationship of age patterns among different age groups
  • By adopting triple translation loss, the progressive mappings of different age domains are fully correlated. The generator is encouraged to be reusable, generating synthesized faces with the more evident aging effect.
  • Enhanced adversarial loss is adopted to effectively model the complex distribution of age domains

Dataset - Cross-Age Celebrity Dataset (CACD)


Keep Thinking!!!

No comments: