REVIEW 3 cited by
DicFace: Dirichlet-Constrained Variational Codebook Learning for Temporally Coherent Video Face Restoration
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
DicFace: Dirichlet-Constrained Variational Codebook Learning for Temporally Coherent Video Face Restoration
read the original abstract
Video face restoration faces a critical challenge in maintaining temporal consistency while recovering fine facial details from degraded inputs. This paper presents a novel approach that extends Vector-Quantized Variational Autoencoders (VQ-VAEs), pretrained on static high-quality portraits, into a video restoration framework through variational latent space modeling. Our key innovation lies in reformulating discrete codebook representations as Dirichlet-distributed continuous variables, enabling probabilistic transitions between facial features across frames. A spatio-temporal Transformer architecture jointly models inter-frame dependencies and predicts latent distributions, while a Laplacian-constrained reconstruction loss combined with perceptual (LPIPS) regularization enhances both pixel accuracy and visual quality. Comprehensive evaluations on blind face restoration, video inpainting, and facial colorization tasks demonstrate state-of-the-art performance. This work establishes an effective paradigm for adapting intensive image priors, pretrained on high-quality images, to video restoration while addressing the critical challenge of flicker artifacts. The source code has been open-sourced and is available at https://github.com/fudan-generative-vision/DicFace.
Forward citations
Cited by 3 Pith papers
-
DVFace: Spatio-Temporal Dual-Prior Diffusion for Video Face Restoration
DVFace uses a spatio-temporal dual-codebook and asymmetric fusion in a one-step diffusion model to deliver better video face restoration quality, temporal consistency, and identity preservation than recent methods.
-
Persistent Free Volume Governs (Anti)plasticization in Chitosan-Water Mixtures
A model of dynamically accessible free volume regions connected through additive-accessible volumes explains the antiplasticization followed by plasticization of elastic properties in chitosan-water mixtures.
-
Persistent Free Volume Governs (Anti)plasticization in Chitosan-Water Mixtures
Dynamically accessible free volume, enabled by connected water-accessible regions, is proposed to govern antiplasticization then plasticization of elastic properties in chitosan–water mixtures.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.