Cross-resolution feature attention network for image super-resolution

Liu, Anqi; Li, Sumei; Chang, Yongli

doi:10.1007/s00371-022-02519-w

Cited by 6 publications

(5 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also provide a comparison of light-weight image SR models. The comparison methods include CARN, 57 FALSR-A, 58 IMDN, 59 LAPAR-A, 60 LatticeNet, 61 SwinIR-light, 19 Swin2SR-s, 22 ESRT, 62 ELAN-light, 21 SPIN, 63 and CRAFT 64 . Regarding the light-weight DiNAT-SR structure, the RDiNAG number N is set to 4.…”

Section: Methodsmentioning

confidence: 99%

Image super-resolution using dilated neighborhood attention transformer

Chen,

Zuo,

et al. 2024

J. Electron. Imag.

View full text Add to dashboard Cite

Transformer-based methods have achieved impressive performance in image super-resolution (SR). To reduce the computational cost and redundancy of global attention, most transformer-based methods adopt a localized attention mechanism, which diminishes the desirable characteristics of self-attention (SA), such as the effective modeling of long-range dependencies and the ability to capture a global receptive field. To alleviate this problem, we propose a dilated neighborhood attention transformer for image SR (DiNAT-SR) to improve SwinIR for image SR; in it, we replace SA with DiNA to capture more global data and allow the receptive field to grow exponentially. In addition, we also introduce a convolutional modulation block into the transformer to enhance the visual representation and facilitate smoother convergence during training. Our research has, for the first time, confirmed the feasibility of DiNA in the field of image SR. Experimental results have demonstrated the effectiveness of DiNAT-SR with better results compared with SwinIR on most benchmarks in terms of both quantitatively and visually. We also provide a comparison of light-weight image SR models, and our model performs better than SwinIR-light on all benchmarks, with similar total numbers of parameters and floating-point operations. The effectiveness of each introduced component is also validated by an ablation study.

show abstract

Section: Methodsmentioning

confidence: 99%

Image super-resolution using dilated neighborhood attention transformer

Chen,

Zuo,

et al. 2024

J. Electron. Imag.

View full text Add to dashboard Cite

show abstract

“…To reduce the impact of background information or distractors, Zhang et al [37] proposed a novel Siamese anchor-free network based on criss-cross attention, obtaining more accurate and robust tracking results. Considering that most attention mechanisms are only processed at a single resolution, Liu et al [38] proposed a cross-resolution feature attention mechanism to progressively reconstruct images at different scale factors. Similarly, we adapt and improve channel and spatial attention mechanisms to enhance reconstruction performance.…”

Section: Attention Mechanismmentioning

confidence: 99%

High-frequency channel attention and contrastive learning for image super-resolution

Yan,

Yin

2024

Vis Comput

View full text Add to dashboard Cite

Over the last decade, convolutional neural networks (CNNs) have allowed remarkable advances in single image super-resolution (SISR). In general, recovering high-frequency features is crucial for high-performance models. High-frequency features suffer more serious damages than low-frequency features during downscaling, making it hard to recover edges and textures. In this paper, we attempt to guide the network to focus more on high-frequency features in restoration from both channel and spatial perspectives. Specifically, we propose a high-frequency channel attention (HFCA) module and a frequency contrastive learning (FCL) loss to aid the process. For the channel-wise perspective, the HFCA module rescales channels by predicting statistical similarity metrics of the feature maps and their high-frequency components. For the spatial perspective, the FCL loss introduces contrastive learning to train a spatial mask that adaptively assigns high-frequency areas with large scaling factors. We incorporate the proposed HFCA module and FCL loss into an EDSR baseline model to construct the proposed lightweight high-frequency channel contrastive network (HFCCN). Extensive experimental results show that it can yield markedly improved or competitive performances compared to the state-of-the-art networks of similar model parameters.

show abstract

“…In recent years, researchers have proposed many lightweight SR algorithms that aim to provide high‐quality reconstruction results while maintaining low computational complexity. For example, enhanced deep SR [26] and residual channel attention networks [27] utilize residual structures and channel attention mechanisms to improve image reconstruction quality. These methods maintain relatively low computational complexity while improving SR performance.…”

Section: Related Workmentioning

confidence: 99%

Image super‐resolution reconstruction based on implicit image functions

Lin,

Yang

2024

IET Image Processing

View full text Add to dashboard Cite

Image super‐resolution (SR) reconstruction is a key technique for improving image quality and details. Conventional methods are frequently limited by interpolation, filtering, or statistical approaches; thus, they are unable to reconstruct high‐quality continuously enlarged images with detailed information. This study proposes an image SR reconstruction network model, called LALNet, based on implicit image functions and residual multilayered perceptron (RAMLP) with an attention mechanism. Through the implicit image function and RAMLP + attention, high‐quality SR reconstruction with continuous scale factors is achieved, and LALNets can run on embedded edge computing platforms. This method exhibits the following advantages: lightweight network structure reduces computing requirements, introduction of implicit image functions and RAMLP improves reconstruction quality, and attention mechanism suppresses artefacts and distortions. Experimental results show that LALNet outperforms traditional and other deep learning methods in terms of reconstruction performance and computational efficiency. This research provides new ideas and methods for the further development of the field of image SR reconstruction.

show abstract

Cross-resolution feature attention network for image super-resolution

Cited by 6 publications

References 47 publications

Image super-resolution using dilated neighborhood attention transformer

Image super-resolution using dilated neighborhood attention transformer

High-frequency channel attention and contrastive learning for image super-resolution

Image super‐resolution reconstruction based on implicit image functions

Contact Info

Product

Resources

About