Efficient Video Face Enhancement with Enhanced Spatial-Temporal Consistency

TL;DR

This paper introduces an efficient blind video face enhancement method that improves quality and temporal consistency of compressed face videos using a novel 3D-VQGAN-based framework with a de-flickering mechanism.

Contribution

The paper proposes a new 3D-VQGAN-based framework with a two-stage learning process and de-flickering for improved face video enhancement, surpassing current methods in efficiency and quality.

Findings

Outperforms state-of-the-art methods on VFHQ-Test dataset

Reduces flickering and improves visual quality of face videos

Achieves faster processing times compared to existing approaches

Abstract

As a very common type of video, face videos often appear in movies, talk shows, live broadcasts, and other scenes. Real-world online videos are often plagued by degradations such as blurring and quantization noise, due to the high compression ratio caused by high communication costs and limited transmission bandwidth. These degradations have a particularly serious impact on face videos because the human visual system is highly sensitive to facial details. Despite the significant advancement in video face enhancement, current methods still suffer from $i)$ long processing time and $ii)$ inconsistent spatial-temporal visual effects (e.g., flickering). This study proposes a novel and efficient blind video face enhancement method to overcome the above two challenges, restoring high-quality videos from their compressed low-quality versions with an effective de-flickering mechanism. In particular, the proposed method develops upon a 3D-VQGAN backbone associated with spatial-temporal codebooks recording high-quality portrait features and residual-based temporal information. We develop a two-stage learning framework for the model. In Stage \Rmnum{1}, we learn the model with a regularizer mitigating the codebook collapse problem. In Stage \Rmnum{2}, we learn two transformers to lookup code from the codebooks and further update the encoder of low-quality videos. Experiments conducted on the VFHQ-Test dataset demonstrate that our method surpasses the current state-of-the-art blind face video restoration and de-flickering methods on both efficiency and effectiveness. Code is available at \url{https://github.com/Dixin-Lab/BFVR-STC}.