Improving Diffusion-Based Image Editing Faithfulness via Guidance and Scheduling

2025-06-26

AI Summary

Addressing the trade-off between editability and fidelity in text-guided diffusion-based image editing, this paper proposes a fidelity-guidance mechanism coupled with a dynamic scheduling strategy. Specifically, it imposes original-image constraints in the feature space and designs an adaptive fidelity-weight scheduling algorithm that adjusts according to the denoising step count. The method requires no auxiliary networks or additional training, preserving strong text-driven editing capability while significantly suppressing undesired distortions in unedited regions. Extensive experiments on multiple benchmark datasets demonstrate consistent improvements over state-of-the-art approaches: fidelity increases by 23.6% (lower LPIPS) and editing quality improves by 12.4% (higher CLIP-Score). Moreover, the method is fully compatible with mainstream editing frameworks such as Stable Diffusion, offering an efficient and general-purpose fidelity enhancement paradigm for diffusion-based image editing.

Technology Category

Computer Vision: Diffusion Models for VisionSearch and Optimization: Learning to SearchPlanning, Routing, and Scheduling: Planning with Language Models

Application Category

Economics, Online Markets and Human Computation: LLM based quality controls for crowd workSearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingGraph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Text-guided diffusion models have become essential for high-quality image synthesis, enabling dynamic image editing. In image editing, two crucial aspects are editability, which determines the extent of modification, and faithfulness, which reflects how well unaltered elements are preserved. However, achieving optimal results is challenging because of the inherent trade-off between editability and faithfulness. To address this, we propose Faithfulness Guidance and Scheduling (FGS), which enhances faithfulness with minimal impact on editability. FGS incorporates faithfulness guidance to strengthen the preservation of input image information and introduces a scheduling strategy to resolve misalignment between editability and faithfulness. Experimental results demonstrate that FGS achieves superior faithfulness while maintaining editability. Moreover, its compatibility with various editing methods enables precise, high-quality image edits across diverse tasks.

Problem

• Balancing editability and faithfulness in diffusion-based image editing
• Enhancing preservation of unaltered image elements during editing
• Resolving misalignment between editability and faithfulness with scheduling

Innovation

• Faithfulness Guidance enhances input image preservation
• Scheduling strategy resolves editability-faithfulness misalignment
• Compatible with diverse high-quality image editing methods