1 paper · 2 filters
Candi Zheng, Yuan Lan
Diffusion models with prompt and reference image-guided editing have seen rapid progress, yet they remain too coarse for pixel-level control. One promising direction is to incorpor…