AI RESEARCH
Spatial Gram Alignment for Ultra-High-Resolution Image Synthesis
arXiv CS.CV
•
ArXi:2605.20808v1 Announce Type: new Modern ultra-high-resolution image synthesis relies heavily on the robust generative capacity of large-scale pre-trained Latent Diffusion Models (LDMs). While recent representation alignment methods have proven effective by distilling visual priors from foundation models (e.g., SAM or DINO) into generative latent features, scaling these approaches to pre-trained LDMs at extreme resolutions exposes a critical learnability-fidelity conflict.