HierEdit: Region-Aware Hierarchical Diffusion for Efficient High-Resolution Editing

2026-05-17

AI Summary

Existing diffusion models incur substantial computational costs in high-resolution image editing and rely heavily on large-scale high-resolution training data, hindering their scalability to ultra-high resolutions such as 4K. This work proposes a region-aware hierarchical diffusion framework that first identifies the target editing region and generates a reference on a low-resolution proxy image, then refines only the corresponding local region in the original high-resolution image using a window-based MMDiT architecture. The unmodified regions of the original image are reused as conditioning context, while the low-resolution proxy provides structural guidance and intermediate denoising supervision. Without requiring specialized high-resolution training data, the method achieves efficient 4K image editing for the first time, significantly accelerating inference while preserving global semantic consistency and attaining visual quality comparable to state-of-the-art approaches on general-resolution datasets.

Abstract

High-resolution image editing is essential for professional and creative applications, yet existing multimodal diffusion-based editors remain computationally inefficient and constrained to relatively low resolutions. Current approaches redundantly process the entire image canvas or rely on large-scale high-resolution datasets, resulting in substantial training and inference costs. We introduce HierEdit, a region-aware hierarchical diffusion framework designed for efficient and scalable high-resolution image editing. Our method first performs edits on a low-resolution proxy using an off-the-shelf editing model to generate a reference and to localize the modified regions. A hierarchical local-window diffusion model (\textbf{Local-Window MMDiT}) that refines only edited regions within the original high-res image, while reusing the unaltered regions as conditioning inputs. The low-resolution proxy further provides structural guidance and intermediate denoising supervision (\textbf{Inference Acceleration}) , ensuring consistent global semantics and stable generation without the need for full-resolution attention computation. This targeted and hierarchical design enables fast, high-fidelity editing of images up to 4K resolution without any specialized high-resolution training data. Extensive experiments demonstrate that HierEdit achieves competitive visual quality on commodity-resolution datasets while significantly accelerating inference and extending seamlessly to ultra-high-resolution 4K editing. Please check our {\href{https://peteryyzhang.github.io/HierEdit-page/}{\textbf{Project Page}}}.

Problem

• high-resolution image editing
• diffusion models
• computational efficiency
• scalability
• multimodal editing

Innovation

• Hierarchical Diffusion
• Region-Aware Editing
• Local-Window MMDiT
• Inference Acceleration
• High-Resolution Image Editing