Saved in:
Bibliographic Details
Main Authors: Liao, Yucheng, Liang, Jiajun, Cui, Kaiqian, Zhao, Baoquan, Xie, Haoran, Liu, Wei, Li, Qing, Mao, Xudong
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.01755
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915879705378816
author Liao, Yucheng
Liang, Jiajun
Cui, Kaiqian
Zhao, Baoquan
Xie, Haoran
Liu, Wei
Li, Qing
Mao, Xudong
author_facet Liao, Yucheng
Liang, Jiajun
Cui, Kaiqian
Zhao, Baoquan
Xie, Haoran
Liu, Wei
Li, Qing
Mao, Xudong
contents Instruction-based image editing through natural language has emerged as a powerful paradigm for intuitive visual manipulation. While recent models achieve impressive results on single edits, they suffer from severe quality degradation under multi-turn editing. Through systematic analysis, we identify progressive loss of high-frequency information as the primary cause of this quality degradation. We present FreqEdit, a training-free framework that enables stable editing across 10+ consecutive iterations. Our approach comprises three synergistic components: (1) high-frequency feature injection from reference velocity fields to preserve fine-grained details, (2) an adaptive injection strategy that spatially modulates injection strength for precise region-specific control, and (3) a path compensation mechanism that periodically recalibrates the editing trajectory to prevent over-constraint. Extensive experiments demonstrate that FreqEdit achieves superior performance in both identity preservation and instruction following compared to seven state-of-the-art baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2512_01755
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FreqEdit: Preserving High-Frequency Features for Robust Multi-Turn Image Editing
Liao, Yucheng
Liang, Jiajun
Cui, Kaiqian
Zhao, Baoquan
Xie, Haoran
Liu, Wei
Li, Qing
Mao, Xudong
Computer Vision and Pattern Recognition
Instruction-based image editing through natural language has emerged as a powerful paradigm for intuitive visual manipulation. While recent models achieve impressive results on single edits, they suffer from severe quality degradation under multi-turn editing. Through systematic analysis, we identify progressive loss of high-frequency information as the primary cause of this quality degradation. We present FreqEdit, a training-free framework that enables stable editing across 10+ consecutive iterations. Our approach comprises three synergistic components: (1) high-frequency feature injection from reference velocity fields to preserve fine-grained details, (2) an adaptive injection strategy that spatially modulates injection strength for precise region-specific control, and (3) a path compensation mechanism that periodically recalibrates the editing trajectory to prevent over-constraint. Extensive experiments demonstrate that FreqEdit achieves superior performance in both identity preservation and instruction following compared to seven state-of-the-art baselines.
title FreqEdit: Preserving High-Frequency Features for Robust Multi-Turn Image Editing
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.01755