SANA I2I: A Text Free Flow Matching Framework for Paired Image to Image Translation with a Case Study in Fetal MRI Artifact Reduction

📰 ArXiv cs.AI

SANA I2I is a text-free framework for paired image-to-image translation using a flow matching approach

advanced Published 2 Apr 2026
Action Steps
  1. Learn a conditional flow-matching model in latent space using paired source-target images
  2. Map a target image distribution to another one using a conditional velocity field
  3. Apply the framework to paired image-to-image translation tasks, such as fetal MRI artifact reduction
  4. Evaluate the performance of SANA I2I on various image translation tasks
Who Needs to Know This

Computer vision engineers and researchers on a team can benefit from SANA I2I for image translation tasks, such as artifact reduction in medical imaging

Key Insight

💡 SANA I2I removes textual conditioning and relies on paired source-target images to learn a conditional flow-matching model

Share This
📸 Introducing SANA I2I: a text-free framework for paired image-to-image translation using flow matching! 💻
Read full paper → ← Back to News