Preprint
Oct 2026
MiMaAlign: Hierarchical Micro-Macro Alignment for Physically Plausible Video Generation
Text-to-video (T2V) diffusion models have made remarkable progress in generating high-fidelity visuals, yet their outputs still often exhibit physically implausible phenomena, including discontinuous dynamics and implausible event interactions. Existing representation alignment methods mainly transfer generic video fea...
Yan-Song Liu, Li-Tong Gong, Wei-Jie Li et al.
· 0 citations