ER-Chord: Spatially Consistent One-Step Image Editing via Equal-Energy Reliability Redistribution
One-step generative models enable fast text-guided image editing, yet they often introduce unintended changes to regions that should remain unchanged.
ProofPaper ↗
Key points
- Temporal aggregation stabilizes editing fields across noise levels, but leaves their spatial allocation insensitive to local variations in reliability.
- Consequently, temporal stability alone does not ensure spatially consistent editing.
- To address this lim- itation, we propose ER-Chord, a training-free and inversion-free framework for reliability-guided spatial redistribution.
- We formulate this allocation as an equal-energy op- timization that redistributes control energy according to local reliability while preserving the original editing directions and total control energy.
Sources (1)
- [1]ER-Chord: Spatially Consistent One-Step Image Editing via Equal-Energy Reliability RedistributionarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 21, 05:53 AM
One-step generative models enable fast text-guided image editing, yet they often introduce unintended changes to regions that should remain unchanged.
Temporal aggregation stabilizes editing fields across noise levels, but leaves their spatial allocation insensitive to local variations in reliability.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 20, 2026WildAIGC-Bench: Can AI Image Detectors Handle the Wild? A Benchmark of Generated and Edited Images in Online Circulation
- Oct 20, 2026MMEnterprise-Bench: Evaluating Enterprise Agents over Evolving Multimodal Records
- Oct 19, 2026AutoRedClaw: Automated Red-Teaming for Claw-Style Agents
- Oct 19, 2026EGI-BENCH: An Enterprise General Intelligence Benchmark Suite
- Oct 19, 2026When Coding Agents Outpace Review: Evaluating PRD-Driven Development in Mature Enterprise Repositories
- Oct 19, 2026VibeTraining-Bench: Data-Trainer-Harness Co-Design for Iterative System Improvement