GAN Inversion on CelebA-HQ (test)
23.59PSNRmGANprior
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| mGANpriorbackbone=PGGAN, latent_codes=multiple2019.12 | 23.59 | 0.4432 | — | — | |
| Encoder initialization for optimizationbackbone=PGGAN, latent_codes=single2019.12 | 20.33 | 0.5321 | — | — | |
| Optimizing a single latent codebackbone=PGGAN, latent_codes=single2019.12 | 19.17 | 0.5797 | — | — | |
| Learning an encoderbackbone=PGGAN, latent_codes=single2019.12 | 11.18 | 0.6992 | — | — | |
| HFGI3DEvaluation Protocol=User study2022.11 | — | — | 90.7 | 92.5 |