FreePCA improves training-free long video generation by projecting global and local attention features into PCA space, selecting consistent appearance components from global and motion components from local, then progressively fusing them.
Videocrafter2: Overcoming data limitations for high-quality video diffu- sion models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
FreePCA: Integrating Consistency Information across Long-short Frames in Training-free Long Video Generation via Principal Component Analysis
FreePCA improves training-free long video generation by projecting global and local attention features into PCA space, selecting consistent appearance components from global and motion components from local, then progressively fusing them.