< Explain other AI papers

Follow-Your-Shape: Shape-Aware Image Editing via Trajectory-Guided Region Control

Zeqian Long, Mingzhe Zheng, Kunyu Feng, Xinhua Zhang, Hongyu Liu, Harry Yang, Linfeng Zhang, Qifeng Chen, Yue Ma

2025-08-12

Follow-Your-Shape: Shape-Aware Image Editing via Trajectory-Guided
  Region Control

Summary

This paper talks about Follow-Your-Shape, a new way to edit images that lets you change the shape of objects very precisely without messing up other parts of the picture. It uses smart techniques to understand exactly where and how to edit, preserving the rest of the image perfectly.

What's the problem?

The problem is that earlier image editing tools often fail when trying to change large shapes in images. They either don't achieve the desired shape changes well or accidentally alter parts of the image that should stay the same, like the background, which lowers the overall quality.

What's the solution?

The researchers created a method called Follow-Your-Shape that uses a Trajectory Divergence Map to compare how the original image and the edited parts move differently over time. This map helps pinpoint exactly where to edit. Then, a Scheduled KV Injection technique uses this information to make sure the edits happen smoothly and accurately only in the target areas, keeping the rest of the image intact. The system works without needing extra training or masks to show what to edit.

Why it matters?

This matters because it allows artists and designers to make big shape changes in images without ruining other important details, making the editing faster, easier, and of higher quality. This can be very useful for creating better visuals in movies, games, advertising, and other creative fields.

Abstract

Follow-Your-Shape framework uses a Trajectory Divergence Map and Scheduled KV Injection to enable precise and controllable shape editing in images while preserving non-target content.