AI Paper Atlas

Research map / Image generation / Editing and style transfer

LayerVerse: Finding the Sweet Spot for KV-Injection in Training-Free Image Editing

Mikhail Zhirnov, Arsen Kuzhamuratov, Andrey Kuznetsov, Ivan Oseledets, Konstantin Sobolev

ECCV 2026 · Image generation › Editing and style transfer · approach: Inversion-based diffusion editing · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
apply LayerVerse layer allocation to FireFlowFireFlow (FLUX.1-dev) · PIE-BenchLPIPS: 64.59 -> 37.82Improved
apply LayerVerse to FireFlowFireFlow (FLUX.1-dev) · PIE-BenchEditScore: 3.60 -> 5.05Improved
apply LayerVerse to FireFlowFireFlow (FLUX.1-dev) · PIE-BenchStructure distance: 9.48 -> 9.36Improved
apply LayerVerse to FireFlowFireFlow (FLUX.1-dev) · PIE-BenchCLIP whole: 23.68 -> 24.39Mixed
apply LayerVerse to RF-SolverRF-Solver (FLUX.1-dev) · PIE-Benchavg rank: 5.88 -> 3.38Improved
use only global KV injection instead of combined (ablation)LayerVerse · PIE-Bench FLUX.1-devStructure distance: 9.36 -> 11.01Hurt
use only masked KV injection instead of combined (ablation)LayerVerse · PIE-Bench FLUX.1-devStructure distance: 9.36 -> 19.03Hurt
use random block selection with same budget (ablation)LayerVerse (FireFlow) · PIE-Bench FLUX.1-devStructure distance: 9.36 -> 60.54Hurt
use random block selection with same budget (ablation)LayerVerse (Euler) · PIE-Bench FLUX.1-devStructure distance: 12.24 -> 64.16Hurt
use quadratic pairwise error model instead of linear (ablation)LayerVerse MILP · PIE-Bench FLUX.1-devreconstruction PSNRImproved
reduce calibration set from 115 to 10 images (ablation)LayerVerse · FLUX.1-dev calibrationcost / performance: ~348 -> ~30 GPU hoursNo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy