1 paper · 1 filter
Zhexiao Xiong, Wei Xiong, Jing Shi +3
Recent approaches in text-to-image customization have primarily focused on preserving the identity of the input subject, but often fail to control the spatial location and size of…