Local-Global Fusion Network for Video Super-Resolution

Su, Da; Wang, Hua; Jin, Longcun; Sun, Xianfang; Peng, Xinyi

doi:10.1109/access.2020.3025780

Cited by 7 publications

(3 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[49]- [51] and recurrent [21], [22], [52]. The sliding window method is to aggregate information from adjacent frames to the reference frame, which is typically the center frame, to obtain the HR results of the reference frame.…”

Section: Related Workmentioning

confidence: 99%

Recurrent Feature Updating Network for Video Super-Resolution

Li,

Zhu

2023

IEEE Access

View full text Add to dashboard Cite

Temporal modeling is the essential to achieve video super-resolution. Most models use alignment or recurrent methods to directly exploit the temporal information of consecutive frames. However, the feature information extracted directly from the input frames is coarse, which affects the performance and generalization ability of the model. Thus, in this paper, we propose to investigate the role of recurrent feature updating networks. The feature update module is proposed to update and optimize the input information to improve the accuracy of features. We design the model as a multi-stage form to improve performance and generalization. Moreover, to further improve the performance of the model, we propose the difference supplement module. It allows the model to exploit the temporal differences between groups to complement the missing information in the output of each stage. Experiments demonstrate that our proposed model achieves 27.78 dB, 30.44 dB, and 39.38 dB on Vid4, SPMCS, and UDM10, respectively, which indicates that our proposed model achieves state-of-the-art performance on video super-resolution. Code is available at https://github.com/gfli123/RFUN.

show abstract

Section: Related Workmentioning

confidence: 99%

Recurrent Feature Updating Network for Video Super-Resolution

Li,

Zhu

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…We selected the following 17 methods: BasicVSR++ [6], COMISR [16], DBVSR [34], EGVSR [5], LGFN [21], RBPN [8], Real-ESRGAN [23], RealSR [11], RSDN [10], SOF-VSR-BD [22], SOF-VSR-BI [22], SwinIR [18], TM-Net [31], VRT [17], ahq-11 [1], amq-12 [1], and bicubic interpolation.…”

Section: B Super-resolution Modelsmentioning

confidence: 99%

Compressed Video Quality Assessment for Super-Resolution: A Benchmark and a Quality Metric

Evgeney¹,

Molodetskikh²,

Vatolin³

et al. 2022

KIAM Prepr.

View full text Add to dashboard Cite

We developed a super-resolution (SR) benchmark to analyze SR capabilities to upscale compressed videos. The dataset for the benchmark was collected using video codecs of 5 different compression standards. We assessed 17 state-of-the-art SR models using our benchmark and evaluated their ability to preserve scene context and their robustness to compression artifacts. To get an accurate perceptual ranking of SR models, we conducted a crowd-sourced side-by-side comparison of SR results. We also analyzed the results of the benchmark and developed an objective quality assessment metric based on existing best-performing objective metrics. Our metric outperforms other video quality metrics by Spearman correlation with subjective scores for the task of upscaling compressed videos.

show abstract

“…Local-Global Fusion Network (LGFN) for Video SR (Su et al, 2020) propose deformable convolutions (DCs) with decreased multi-dilation convolution units (DMDCUs) for efficient implicit alignment. Moreover, a structure with two branches, consisting of a Local Fusion Module (LFM) and a Global Fusion Module (GFM), is proposed to combine information from two different aspects.…”

Section: Convnet Architectures For Feature Fusionmentioning

confidence: 99%

Deep Learning for Image/Video Restoration and Super-resolution

Tekalp¹

2022

View full text Add to dashboard Cite

Recent advances in neural signal processing led to significant improvements in the performance of learned image/video restoration and super-resolution (SR). An important benefit of data-driven deep learning approach to image processing is that neural models can be optimized for any differentiable loss function, including perceptual loss functions, leading to perceptual image/video restoration and SR, which cannot be easily handled by traditional model-based methods.We start with a brief problem statement and a short discussion on traditional vs. data-driven solutions. We next review recent advances in neural architectures, such as residual blocks, dense connections, residual-in-residual dense blocks, residual blocks with generative neurons, self-attention and visual transformers. We then discuss loss functions and evaluation (assessment) criteria for image/video restoration and SR, including fidelity (distortion) and perceptual criteria, and the relation between them, where we briefly review the perception vs. distortion trade-off.We can consider learned image/video restoration and SR as learning either a nonlinear regressive mapping from degraded to ideal images based on the universal approximation theorem, or a generative model that captures the probability distribution of ideal images. We first review regressive

show abstract

Local-Global Fusion Network for Video Super-Resolution

Cited by 7 publications

References 46 publications

Recurrent Feature Updating Network for Video Super-Resolution

Recurrent Feature Updating Network for Video Super-Resolution

Compressed Video Quality Assessment for Super-Resolution: A Benchmark and a Quality Metric

Deep Learning for Image/Video Restoration and Super-resolution

Contact Info

Product

Resources

About