AI Resonance · 阅读最新日报 · AI 入门推荐

2026-10-23 · America/Los_Angeles · 论文 · #4

Editing Images into Reusable Layers via Compositional Edit-Prior Distillation

High-quality image editing requires precise modification of target regions while preserving the content of unrelated regions. Existing editing models produce flattened outputs that limit precise local manipulation and layer reuse, while decomposition methods typically optimize layer recovery separately from instruction-driven editing. To this end, we introduce LayerEditor, a unified framework that couples instruction-guided image editing with controllable generation of reusable RGBA layers. At its core, In-Context RGB-Alpha Token Modeling jointly represents the edited canvas and an ordered…

Editing Images into Reusable Layers via Compositional Edit-Prior Distillation

热度 55.5 / 100;排名与评分保留该期记录。