By the same authors

From the same journal

From the same journal

Single Image Super Resolution via Neighbor Reconstruction

Research output: Contribution to journalArticle

Standard

Single Image Super Resolution via Neighbor Reconstruction. / Zhang, Zhihong; Xu, Chen; Zhang, Zhonghao; Chen, Guo; Cai, Yide; Wang, Zeli; Li, Heng; Hancock, Edwin R.

In: Pattern Recognition Letters, 22.04.2019.

Research output: Contribution to journalArticle

Harvard

Zhang, Z, Xu, C, Zhang, Z, Chen, G, Cai, Y, Wang, Z, Li, H & Hancock, ER 2019, 'Single Image Super Resolution via Neighbor Reconstruction', Pattern Recognition Letters. https://doi.org/10.1016/j.patrec.2019.04.021

APA

Zhang, Z., Xu, C., Zhang, Z., Chen, G., Cai, Y., Wang, Z., ... Hancock, E. R. (2019). Single Image Super Resolution via Neighbor Reconstruction. Pattern Recognition Letters. https://doi.org/10.1016/j.patrec.2019.04.021

Vancouver

Zhang Z, Xu C, Zhang Z, Chen G, Cai Y, Wang Z et al. Single Image Super Resolution via Neighbor Reconstruction. Pattern Recognition Letters. 2019 Apr 22. https://doi.org/10.1016/j.patrec.2019.04.021

Author

Zhang, Zhihong ; Xu, Chen ; Zhang, Zhonghao ; Chen, Guo ; Cai, Yide ; Wang, Zeli ; Li, Heng ; Hancock, Edwin R. / Single Image Super Resolution via Neighbor Reconstruction. In: Pattern Recognition Letters. 2019.

Bibtex - Download

@article{e71341de45c94386891de5624dee2fbe,
title = "Single Image Super Resolution via Neighbor Reconstruction",
abstract = "Super Resolution (SR) is a complex, ill-posed problem where the aim is to construct the mapping between the low and high resolution manifolds of image patches. Anchored neighborhood regression for SR (namely A+  [27]) has shown promising results. In this paper we present a new regression-based SR algorithm that overcomes the limitations of A+ and benefits from an innovative and simple Neighbor Reconstruction Method (NRM). This is achieved by vector operations on an anchored point and its corresponding neighborhood. NRM reconstructs new patches which are closer to the anchor point in the manifold space. Our method is robust to NRM sparsely-sampled points: increasing PSNR by 0.5 dB compared to the next best method. We comprehensively validate our technique on standardised datasets and compare favourably with the state-of-the-art methods: we obtain PSNR improvement of up to 0.21 dB compared to previously-reported work.",
keywords = "Manifold learning, Neighbor Reconstruction, Super Resolution",
author = "Zhihong Zhang and Chen Xu and Zhonghao Zhang and Guo Chen and Yide Cai and Zeli Wang and Heng Li and Hancock, {Edwin R.}",
note = "This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy.",
year = "2019",
month = "4",
day = "22",
doi = "10.1016/j.patrec.2019.04.021",
language = "Undefined/Unknown",
journal = "Pattern Recognition Letters",
issn = "0167-8655",
publisher = "Elsevier",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Single Image Super Resolution via Neighbor Reconstruction

AU - Zhang, Zhihong

AU - Xu, Chen

AU - Zhang, Zhonghao

AU - Chen, Guo

AU - Cai, Yide

AU - Wang, Zeli

AU - Li, Heng

AU - Hancock, Edwin R.

N1 - This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy.

PY - 2019/4/22

Y1 - 2019/4/22

N2 - Super Resolution (SR) is a complex, ill-posed problem where the aim is to construct the mapping between the low and high resolution manifolds of image patches. Anchored neighborhood regression for SR (namely A+  [27]) has shown promising results. In this paper we present a new regression-based SR algorithm that overcomes the limitations of A+ and benefits from an innovative and simple Neighbor Reconstruction Method (NRM). This is achieved by vector operations on an anchored point and its corresponding neighborhood. NRM reconstructs new patches which are closer to the anchor point in the manifold space. Our method is robust to NRM sparsely-sampled points: increasing PSNR by 0.5 dB compared to the next best method. We comprehensively validate our technique on standardised datasets and compare favourably with the state-of-the-art methods: we obtain PSNR improvement of up to 0.21 dB compared to previously-reported work.

AB - Super Resolution (SR) is a complex, ill-posed problem where the aim is to construct the mapping between the low and high resolution manifolds of image patches. Anchored neighborhood regression for SR (namely A+  [27]) has shown promising results. In this paper we present a new regression-based SR algorithm that overcomes the limitations of A+ and benefits from an innovative and simple Neighbor Reconstruction Method (NRM). This is achieved by vector operations on an anchored point and its corresponding neighborhood. NRM reconstructs new patches which are closer to the anchor point in the manifold space. Our method is robust to NRM sparsely-sampled points: increasing PSNR by 0.5 dB compared to the next best method. We comprehensively validate our technique on standardised datasets and compare favourably with the state-of-the-art methods: we obtain PSNR improvement of up to 0.21 dB compared to previously-reported work.

KW - Manifold learning, Neighbor Reconstruction, Super Resolution

U2 - 10.1016/j.patrec.2019.04.021

DO - 10.1016/j.patrec.2019.04.021

M3 - Article

JO - Pattern Recognition Letters

T2 - Pattern Recognition Letters

JF - Pattern Recognition Letters

SN - 0167-8655

ER -