Posterior Distillation Sampling

arXiv:2311.13831v2 Announce Type: replace
Abstract: We introduce Posterior Distillation Sampling (PDS), a novel optimization method for parametric image editing based on diffusion models. Existing optimization-based methods, which leverage the powerful 2D prior of diffusion models to handle various parametric images, have mainly focused on generation. Unlike generation, editing requires a balance between conforming to the target attribute and preserving the identity of the source content. Recent 2D image editing methods have achieved this balance by leveraging the stochastic latent encoded …

abstract arxiv balance cs.cv diffusion diffusion models distillation editing image images novel optimization parametric posterior prior sampling type