Review Article
Generative Adversarial Network Technologies and Applications in Computer Vision
| Stages | Stage 1 | Stage 2 | Stage 3 |
| Time | 2014.06–2015.11 | 2015.11–2017.01 | 2017.1-today | GAN models | GAN- > DCGAN | DCGAN- > WGAN | WGAN- > today | Improvements | GAN is the beginning of generating the adversarial model | DCGAN uses many new methods to make the model more stable such as batchnorm, ReLU, and leaky ReLU | WGAN uses weight clipping solving the problem of gradient disappearance |
|
|