Mathematical Analysis of DCN-Based Super-Resolution

Lee, Chulhee; Park, J.; Woo, Seongyoun; Kim, Jongyoung; Yoon, Jong-Su

doi:10.1109/access.2020.2992483

Cited by 2 publications

(4 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [20], it was shown that a convolution layer followed by a ReLU layer can be expressed by matrix operations. The patch was expressed as a vector (N x 1) with 2 NK = .…”

Section: A Dynamic Linear Transformationmentioning

confidence: 99%

“…The patch was expressed as a vector (N x 1) with 2 NK = . If there are 64 filters in the convolution layer, the convolution layer followed by the ReLU layer was expressed as follows [20]:…”

Section: A Dynamic Linear Transformationmentioning

confidence: 99%

“…Thus the receptive field is 41 x 41. In [20], it was shown that an output pixel can be expressed as follows:…”

Section: J Nnmentioning

confidence: 99%

See 2 more Smart Citations

One-to-One Mapping-Like Properties of DCN-Based Super-Resolution and its Applicability to Real-World Images

et al. 2021

Self Cite

View full text Add to dashboard Cite

Although super-resolution techniques based on deep neural networks (SRDNN) have drawn significant interest and numerous algorithms have been proposed, they still have reliability problems and produce artefacts when applied to new datasets. In this paper, the working mechanisms of SRDNN techniques are analyzed in terms of data mapping. Since most SRDNN techniques can be viewed as dynamic linear projections, we analyzed a large number of projection vectors (over 70 million) and found that the SRDNN method performs one-to-one mapping-like operations and may be vulnerable to unknown data patterns. Then, we applied several SRDNN techniques to real-world images and analyzed the output images. The current SRDNN methods failed to distinguish the blurred edges/lines due to low resolutions from coding artefacts and enhanced both, even though the SRDNN methods were trained using compressed low-resolution (LR) images. These analyses and results indicate that current SRDNN methods may not be able to provide robust performance and new structures may be necessary for reliable super-resolution performance.

show abstract

“…In [20], it was shown that a convolution layer followed by a ReLU layer can be expressed by matrix operations. The patch was expressed as a vector (N x 1) with 2 NK = .…”

Section: A Dynamic Linear Transformationmentioning

confidence: 99%

“…The patch was expressed as a vector (N x 1) with 2 NK = . If there are 64 filters in the convolution layer, the convolution layer followed by the ReLU layer was expressed as follows [20]:…”

Section: A Dynamic Linear Transformationmentioning

confidence: 99%

See 1 more Smart Citation

One-to-One Mapping-Like Properties of DCN-Based Super-Resolution and its Applicability to Real-World Images

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recently, deep convolutional neural networks (DCN) have been used for many SR problems and they have shown impressive performance improvement. Although the current DCN-based SR techniques have some limitations and reliability issues [44], they can provide noticeably improved performance compared to traditional SR methods under controlled conditions.…”

Section: Introductionmentioning

confidence: 99%

Reliable Perceptual Loss Computation for GAN-Based Super-Resolution With Edge Texture Metric

Kim

Lee

2021

IEEE Access

Self Cite

View full text Add to dashboard Cite

Super-resolution (SR) is an ill-posed problem. Generating high-resolution (HR) images from low-resolution (LR) images remains a major challenge. Recently, SR methods based on deep convolutional neural networks (DCN) have been developed with impressive performance improvement. DCN-based SR techniques can be largely divided into peak signal-to-noise ratio (PSNR)-oriented SR networks and generative adversarial networks (GAN)-based SR networks. In most current GAN-based SR networks, the perceptual loss is computed from the feature maps of a single layer or several fixed layers using a differentiable feature extractor such as VGG. This limited layer utilization may produce overly textured artifacts. In this paper, a new edge texture metric (ETM) is proposed to quantify the characteristics of images and then it is utilized only in the training phase to select an appropriate layer when calculating the perceptual loss. We present experimental results showing that the GAN-based SR network trained with the proposed method achieves qualitative and quantitative perceptual quality improvements compared to many of the existing methods. INDEX TERMSArtificial neural networks, computer vision, image enhancement, image resolution VOLUME 4, 2016This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.

show abstract

Mathematical Analysis of DCN-Based Super-Resolution

Cited by 2 publications

References 26 publications

One-to-One Mapping-Like Properties of DCN-Based Super-Resolution and its Applicability to Real-World Images

One-to-One Mapping-Like Properties of DCN-Based Super-Resolution and its Applicability to Real-World Images

Reliable Perceptual Loss Computation for GAN-Based Super-Resolution With Edge Texture Metric

Contact Info

Product

Resources

About