Skip to content

Source paper

SpotEdit: Selective Region Editing in Diffusion Transformers

Zhibin Qin, Zhenxiong Tan, Zeqing Wang and 2 more

Zhibin Qin, Zhenxiong Tan, Zeqing Wang, Songhua Liu, Xinchao Wang

arXiv 2512.22323 · 202527 upvotes on Hugging Face

Research ideas from this paper (3)

Generated by AI. The paper’s authors did not write these.

AI summary

This paper introduces SpotEdit, a training-free framework designed to accelerate instruction-based image editing in Diffusion Transformers (DiTs). By dynamically identifying non-edited regions using a perceptual similarity metric and selectively reusing cached features for those areas, SpotEdit avoids redundant computation and preserves background fidelity.

How it works (AI summary)

The model uses 'SpotSelector', which calculates an LPIPS-like score from VAE decoder features to classify tokens as edited or non-edited during the diffusion process. It then employs 'SpotFusion', a mechanism that blends cached features from non-edited regions with condition image features using a time-dependent interpolation, allowing the model to skip DiT block computations for stable regions.