Jul 2026
Appearance Pointers - Multimodal Region Control of Diffusion Transformers
This approach introduces the first modality-agnostic interface for localized multimodal control in a DiT without retraining the base model from scratch, offering a simple and extensible path toward precise, region-aware, multimodal guidance in generative image synthesis.
Rahul Sajnani, Yulia Gryaditskaya, R. Měch et al.
· arXiv.org · 0 citations