Diffusion transformer latent spaces are shown to be semantically disentangled, and prompt-difference directions plus a score-distillation step enable zero-shot fine-grained image editing.
smile" in the unseenS′ will be equivalent to directly changing the real image semantic inS with the direction of
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Latent Space Disentanglement in Diffusion Transformers Enables Precise Zero-shot Semantic Editing
Diffusion transformer latent spaces are shown to be semantically disentangled, and prompt-difference directions plus a score-distillation step enable zero-shot fine-grained image editing.