Leveraging Bitstream Metadata for Fast, Accurate, Generalized Compressed Video Quality Enhancement

TL;DR

This paper introduces a deep learning model that uses bitstream metadata to enhance compressed video quality efficiently, handling various compression settings with a single model and outperforming prior methods in accuracy and throughput.

Contribution

We propose a novel deep learning architecture that leverages bitstream metadata for fast, accurate, and generalized video quality enhancement across multiple compression settings.

Findings

Improves restoration accuracy over prior methods

Achieves higher throughput than recent deep-learning-based compression methods

Handles various compression qualities with a single model

Abstract

Video compression is a central feature of the modern internet powering technologies from social media to video conferencing. While video compression continues to mature, for many compression settings, quality loss is still noticeable. These settings nevertheless have important applications to the efficient transmission of videos over bandwidth constrained or otherwise unstable connections. In this work, we develop a deep learning architecture capable of restoring detail to compressed videos which leverages the underlying structure and motion information embedded in the video bitstream. We show that this improves restoration accuracy compared to prior compression correction methods and is competitive when compared with recent deep-learning-based video compression methods on rate-distortion while achieving higher throughput. Furthermore, we condition our model on quantization data which is readily available in the bitstream. This allows our single model to handle a variety of different compression quality settings which required an ensemble of models in prior work.