COLLAR: Cascaded Object-Level Latent Refinement for High-Fidelity Conditional Generation 事件

COLLAR: Cascaded Object-Level Latent Refinement for High-Fidelity Conditional Generation arXiv:2606.00954v1 Announce Type: new Abstract: Achieving high-fidelity object-level control in Diffusion Transformers remains a significant challenge despite the introduction of structural priors like depth and Canny maps. Current object-level conditional generation methods frequently suffer from visual artifacts and struggle to maintain precise control over objects within small localized regions. To addre