SearchChord improves one-step image editing with smarter prompt selection
SearchChord: Prompt Preselection and Field Recomposition for One-Step Image Editing
Computer Vision and Pattern Recognition
Summary
Editing images by giving computers text instructions can be tricky because you want the change to look right, keep parts of the original image, and still run quickly. The people behind SearchChord made a new method that picks good text instructions and pictures to mix together, so the edited image better matches what you want. They use a smart way to choose and combine these instructions without needing extra training. Their tests show that picking the best instructions helps follow the text better, while mixing images keeps more of the original parts.
What this means in practice
- •For digital artists: Improve image editing tools to produce better-targeted changes while preserving original details using prompt preselection and recomposition techniques.
- •For agricultural technology teams: Enhance image editing in agriculture by balancing edit accuracy and preserving key visual content for tasks like crop and weed image adjustments.
Authors
Boyang Deng, Yuzhen Lu
Abstract
Finding effective prompts and seeds for one-step image editors requires balancing edit fidelity, source preservation, and computational cost. To address this challenge, we explore a large collection of candidate edits, where each candidate is defined by a unique prompt-seed pair, and generate corresponding images on demand. We present SearchChord, a training-free framework that leverages a large pool of such candidates, combining text-based preselection, image verification, and latent-field recomposition. Our ablation studies across pool sizes up to 200 candidates show that a pool of 100 offers a cost-effective operating point. Compared with feeding all 100 candidates into recomposition, TextSim preselects a subset of 10 candidates. Rather than uniform averaging, we incorporate image-verifier scores to weight each candidate during recomposition. Experiments on PIE-Bench and the proposed CropWeed-EditBench reveal a trade-off: candidate selection improves target-text alignment, while recomposition better preserves unmodified source content. The oracle verifier achieves the best performance in both preservation and target-text alignment. Practical verification partially recovers semantic alignment, but recomposition still compromises edit fidelity, including in agricultural image edits. These results motivate text-first preselection and stronger verification to balance preservation, fidelity, and cost. Code and data will be made publicly available.