A diffusion-based video enhancement framework uses codec residuals, motion vectors, and QP-conditioned text prompts to improve the perceptual quality of heavily compressed videos.
Compression artifacts reduction by a deep convo- lutional network
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DiffCVE: Diffusion-based Compressed Video Enhancement
A diffusion-based video enhancement framework uses codec residuals, motion vectors, and QP-conditioned text prompts to improve the perceptual quality of heavily compressed videos.