ZONE: Zero-Shot Instruction-Guided Local Editing

TL;DR

ZONE introduces a zero-shot, instruction-guided local image editing method that enables precise, user-friendly modifications of specific regions in images without affecting the rest, outperforming existing techniques.

Contribution

The paper presents a novel zero-shot local editing approach using instruction conversion, region extraction, and seamless blending, improving precision and usability over prior methods.

Findings

Achieves superior local editing quality compared to state-of-the-art methods.

Enables arbitrary region manipulation with a single instruction.

Demonstrates high user-friendliness and editing accuracy.

Abstract

Recent advances in vision-language models like Stable Diffusion have shown remarkable power in creative image synthesis and editing.However, most existing text-to-image editing methods encounter two obstacles: First, the text prompt needs to be carefully crafted to achieve good results, which is not intuitive or user-friendly. Second, they are insensitive to local edits and can irreversibly affect non-edited regions, leaving obvious editing traces. To tackle these problems, we propose a Zero-shot instructiON-guided local image Editing approach, termed ZONE. We first convert the editing intent from the user-provided instruction (e.g., "make his tie blue") into specific image editing regions through InstructPix2Pix. We then propose a Region-IoU scheme for precise image layer extraction from an off-the-shelf segment model. We further develop an edge smoother based on FFT for seamless blending between the layer and the image.Our method allows for arbitrary manipulation of a specific region with a single instruction while preserving the rest. Extensive experiments demonstrate that our ZONE achieves remarkable local editing results and user-friendliness, outperforming state-of-the-art methods. Code is available at https://github.com/lsl001006/ZONE.