Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

VibeEdit edits images via canvas marks instead of text prompts

AI summary

An arXiv paper introduces VibeEdit, an interface that lets users place spatial marks with optional short notes directly on an image; together these annotations form a "canvas instruction" specifying where to edit and what to change, with no separate text prompt. Built on Qwen-Image-Edit with layer-decoupled conditioning that encodes source images and canvas instructions separately, it is trained with region-weighted supervised fine-tuning followed by rubric-guided reinforcement learning on 1.55 million source-target edit pairs with object masks and structured edit descriptions. On the authors' independently constructed, human-curated benchmark of 419 cases, VibeEdit is reported to reach a VLM rubric score of 79.9 and outside-region PSNR of 32.8 dB, versus 67.4 and 24.0 dB for FireRed, the top text-instructed baseline in their evaluation.

In-depth insight

It tests whether pointing and marking, rather than describing, can make target selection easier when several objects look alike.

Computer Science > Computer Vision and Pattern Recognition

arXiv:2610.12229 (cs)

[Submitted on 8 Oct 2026]

Title:VibeEdit: Image Editing with Canvas Instructions

Authors:Jinjing Zhao, Fangyun Wei, Yitong Wang, Xiuyu Wu, Yunuo Chen, Yang Yue, Sirui Zhang, Wenbo Wang, Hongyang Zhang, Dong Chen, Yan Lu, Chang Xu

View a PDF of the paper titled VibeEdit: Image Editing with Canvas Instructions, by Jinjing Zhao and 11 other authors

#VibeEdit#Qwen-Image-Edit#FireRed

Original · 3 min readRead the original
Group chat

No comments yet. Start the conversation.