VeloEdit: Training-Free Consistent and Continuous Instruction-Based Image Editing via Velocity Field Decomposition

Documenti analoghi

• VeloGauss: Learning Physically Consistent Gaussian Velocity Fields from Videos
  di: Lu, Nengbo, et al.
  Pubblicazione: (2026)
• Three Creates All: You Only Sample 3 Steps
  di: Cai, Yuren, et al.
  Pubblicazione: (2026)
• FlowSlider: Training-Free Continuous Image Editing via Fidelity-Steering Decomposition
  di: Endo, Taichi, et al.
  Pubblicazione: (2026)
• SliderEdit: Continuous Image Editing with Fine-Grained Instruction Control
  di: Zarei, Arman, et al.
  Pubblicazione: (2025)
• CoCoEdit: Content-Consistent Image Editing via Region Regularized Reinforcement Learning
  di: Wu, Yuhui, et al.
  Pubblicazione: (2026)