Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhu, Qiang, Zhang, Fan, Chen, Feiyu, Zhu, Shuyuan, Bull, David, Zeng, Bing
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2502.06431
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915947922587648
author Zhu, Qiang
Zhang, Fan
Chen, Feiyu
Zhu, Shuyuan
Bull, David
Zeng, Bing
author_facet Zhu, Qiang
Zhang, Fan
Chen, Feiyu
Zhu, Shuyuan
Bull, David
Zeng, Bing
contents Compressed video super-resolution (SR) aims to generate high-resolution (HR) videos from the corresponding low-resolution (LR) compressed videos. Recently, some compressed video SR methods attempt to exploit the spatio-temporal information in the frequency domain, showing great promise in super-resolution performance. However, these methods do not differentiate various frequency subbands spatially or capture the temporal frequency dynamics, potentially leading to suboptimal results. In this paper, we propose a deep frequency-based compressed video SR model (FCVSR) consisting of a motion-guided adaptive alignment (MGAA) network and a multi-frequency feature refinement (MFFR) module. Additionally, a frequency-aware contrastive loss is proposed for training FCVSR, in order to reconstruct finer spatial details. The proposed model has been evaluated on three public compressed video super-resolution datasets, with results demonstrating its effectiveness when compared to existing works in terms of super-resolution performance (up to a 0.14dB gain in PSNR over the second-best model) and complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2502_06431
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FCVSR: A Frequency-aware Method for Compressed Video Super-Resolution
Zhu, Qiang
Zhang, Fan
Chen, Feiyu
Zhu, Shuyuan
Bull, David
Zeng, Bing
Computer Vision and Pattern Recognition
Compressed video super-resolution (SR) aims to generate high-resolution (HR) videos from the corresponding low-resolution (LR) compressed videos. Recently, some compressed video SR methods attempt to exploit the spatio-temporal information in the frequency domain, showing great promise in super-resolution performance. However, these methods do not differentiate various frequency subbands spatially or capture the temporal frequency dynamics, potentially leading to suboptimal results. In this paper, we propose a deep frequency-based compressed video SR model (FCVSR) consisting of a motion-guided adaptive alignment (MGAA) network and a multi-frequency feature refinement (MFFR) module. Additionally, a frequency-aware contrastive loss is proposed for training FCVSR, in order to reconstruct finer spatial details. The proposed model has been evaluated on three public compressed video super-resolution datasets, with results demonstrating its effectiveness when compared to existing works in terms of super-resolution performance (up to a 0.14dB gain in PSNR over the second-best model) and complexity.
title FCVSR: A Frequency-aware Method for Compressed Video Super-Resolution
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.06431