ICML 2022spotlight7 citations

Improved StyleGAN-v2 based Inversion for Out-of-Distribution Images

Rakshith Subramanyam, Vivek Narayanaswamy, Mark Naufel, Andreas Spanias, Jayaraman J. Thiagarajan

Abstract

Inverting an image onto the latent space of pre-trained generators, e.g., StyleGAN-v2, has emerged as a popular strategy to leverage strong image priors for ill-posed restoration. Several studies have showed that this approach is effective at inverting images similar to the data used for training. However, with out-of-distribution (OOD) data that the generator has not been exposed to, existing inversion techniques produce sub-optimal results. In this paper, we propose SPHInX (StyleGAN with Projection Heads for Inverting X), an approach for accurately embedding OOD images onto the StyleGAN latent space. SPHInX optimizes a style projection head using a novel training strategy that imposes a vicinal regularization in the StyleGAN latent space. To further enhance OOD inversion, SPHInX can additionally optimize a content projection head and noise variables in every layer. Our empirical studies on a suite of OOD data show that, in addition to producing higher quality reconstructions over the state-of-the-art inversion techniques, SPHInX is effective for ill-posed restoration tasks while offering semantic editing capabilities.

BibTeX
@InProceedings{pmlr-v162-subramanyam22a,
  title = 	 {Improved {S}tyle{GAN}-v2 based Inversion for Out-of-Distribution Images},
  author =       {Subramanyam, Rakshith and Narayanaswamy, Vivek and Naufel, Mark and Spanias, Andreas and Thiagarajan, Jayaraman J.},
  booktitle = 	 {Proceedings of the 39th International Conference on Machine Learning},
  pages = 	 {20625--20639},
  year = 	 {2022},
  editor = 	 {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan},
  volume = 	 {162},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {17--23 Jul},
  publisher =    {PMLR},
  pdf = 	 {https://proceedings.mlr.press/v162/subramanyam22a/subramanyam22a.pdf},
  url = 	 {https://proceedings.mlr.press/v162/subramanyam22a.html},
  abstract = 	 {Inverting an image onto the latent space of pre-trained generators, e.g., StyleGAN-v2, has emerged as a popular strategy to leverage strong image priors for ill-posed restoration. Several studies have showed that this approach is effective at inverting images similar to the data used for training. However, with out-of-distribution (OOD) data that the generator has not been exposed to, existing inversion techniques produce sub-optimal results. In this paper, we propose SPHInX (StyleGAN with Projection Heads for Inverting X), an approach for accurately embedding OOD images onto the StyleGAN latent space. SPHInX optimizes a style projection head using a novel training strategy that imposes a vicinal regularization in the StyleGAN latent space. To further enhance OOD inversion, SPHInX can additionally optimize a content projection head and noise variables in every layer. Our empirical studies on a suite of OOD data show that, in addition to producing higher quality reconstructions over the state-of-the-art inversion techniques, SPHInX is effective for ill-posed restoration tasks while offering semantic editing capabilities.}
}
Improved StyleGAN-v2 based Inversion for Out-of-Distribution Images · ICML 2022