Core Concepts
Utilizing coding priors improves compressed video quality.
Abstract
The article introduces the Coding Priors-Guided Aggregation (CPGA) network for enhancing compressed video quality by utilizing coding priors. It addresses the importance of coding information in improving video quality and introduces a new dataset, Video Coding Priors (VCP), to facilitate research in Video Quality Enhancement (VQE). The CPGA network consists of modules that aggregate temporal and spatial information from coding priors, leading to superior performance compared to existing methods. Experimental results demonstrate the effectiveness of leveraging coding priors in enhancing video quality.
Stats
"Experimental results demonstrate the superiority of our method compared to existing state-of-the-art methods."
"Our method achieves a performance gain of more than 0.03dB compared to previous state-of-the-art methods."
"Our model improves about 0.02dB-0.05dB in terms of average ∆PSNR under RA configuration."
Quotes
"Our CPGA outperforms previous methods on public testing sequences."
"Our model achieves an overall performance gain of 0.13dB in terms of ∆PSNR."