LAPRAN

A scalable laplacian pyramid reconstructive adversarial network for flexible compressive sensing reconstruction

Kai Xu, Zhikang Zhang, Fengbo Ren

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper addresses the single-image compressive sensing (CS) and reconstruction problem. We propose a scalable Laplacian pyramid reconstructive adversarial network (LAPRAN) that enables high-fidelity, flexible and fast CS images reconstruction. LAPRAN progressively reconstructs an image following the concept of the Laplacian pyramid through multiple stages of reconstructive adversarial networks (RANs). At each pyramid level, CS measurements are fused with a contextual latent vector to generate a high-frequency image residual. Consequently, LAPRAN can produce hierarchies of reconstructed images and each with an incremental resolution and improved quality. The scalable pyramid structure of LAPRAN enables high-fidelity CS reconstruction with a flexible resolution that is adaptive to a wide range of compression ratios (CRs), which is infeasible with existing methods. Experimental results on multiple public datasets show that LAPRAN offers an average 7.47 dB and 5.98 dB PSNR, and an average 57.93 % and 33.20 % SSIM improvement compared to model-based and data-driven baselines, respectively.

Original languageEnglish (US)
Title of host publicationComputer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings
EditorsMartial Hebert, Vittorio Ferrari, Cristian Sminchisescu, Yair Weiss
PublisherSpringer Verlag
Pages491-507
Number of pages17
ISBN (Print)9783030012489
DOIs
StatePublished - Jan 1 2018
Event15th European Conference on Computer Vision, ECCV 2018 - Munich, Germany
Duration: Sep 8 2018Sep 14 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11214 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other15th European Conference on Computer Vision, ECCV 2018
CountryGermany
CityMunich
Period9/8/189/14/18

Fingerprint

Compressive Sensing
Pyramid
Image reconstruction
Fidelity
Image Reconstruction
Data-driven
Baseline
Compression
Model-based
Experimental Results

Keywords

  • Compressive sensing
  • Feature fusion
  • Laplacian pyramid
  • Reconstruction
  • Reconstructive adversarial network

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Xu, K., Zhang, Z., & Ren, F. (2018). LAPRAN: A scalable laplacian pyramid reconstructive adversarial network for flexible compressive sensing reconstruction. In M. Hebert, V. Ferrari, C. Sminchisescu, & Y. Weiss (Eds.), Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings (pp. 491-507). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11214 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-01249-6_30

LAPRAN : A scalable laplacian pyramid reconstructive adversarial network for flexible compressive sensing reconstruction. / Xu, Kai; Zhang, Zhikang; Ren, Fengbo.

Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings. ed. / Martial Hebert; Vittorio Ferrari; Cristian Sminchisescu; Yair Weiss. Springer Verlag, 2018. p. 491-507 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11214 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Xu, K, Zhang, Z & Ren, F 2018, LAPRAN: A scalable laplacian pyramid reconstructive adversarial network for flexible compressive sensing reconstruction. in M Hebert, V Ferrari, C Sminchisescu & Y Weiss (eds), Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11214 LNCS, Springer Verlag, pp. 491-507, 15th European Conference on Computer Vision, ECCV 2018, Munich, Germany, 9/8/18. https://doi.org/10.1007/978-3-030-01249-6_30
Xu K, Zhang Z, Ren F. LAPRAN: A scalable laplacian pyramid reconstructive adversarial network for flexible compressive sensing reconstruction. In Hebert M, Ferrari V, Sminchisescu C, Weiss Y, editors, Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings. Springer Verlag. 2018. p. 491-507. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-01249-6_30
Xu, Kai ; Zhang, Zhikang ; Ren, Fengbo. / LAPRAN : A scalable laplacian pyramid reconstructive adversarial network for flexible compressive sensing reconstruction. Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings. editor / Martial Hebert ; Vittorio Ferrari ; Cristian Sminchisescu ; Yair Weiss. Springer Verlag, 2018. pp. 491-507 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{ef19bdc110ef45af80bc1746d844c09b,
title = "LAPRAN: A scalable laplacian pyramid reconstructive adversarial network for flexible compressive sensing reconstruction",
abstract = "This paper addresses the single-image compressive sensing (CS) and reconstruction problem. We propose a scalable Laplacian pyramid reconstructive adversarial network (LAPRAN) that enables high-fidelity, flexible and fast CS images reconstruction. LAPRAN progressively reconstructs an image following the concept of the Laplacian pyramid through multiple stages of reconstructive adversarial networks (RANs). At each pyramid level, CS measurements are fused with a contextual latent vector to generate a high-frequency image residual. Consequently, LAPRAN can produce hierarchies of reconstructed images and each with an incremental resolution and improved quality. The scalable pyramid structure of LAPRAN enables high-fidelity CS reconstruction with a flexible resolution that is adaptive to a wide range of compression ratios (CRs), which is infeasible with existing methods. Experimental results on multiple public datasets show that LAPRAN offers an average 7.47 dB and 5.98 dB PSNR, and an average 57.93 {\%} and 33.20 {\%} SSIM improvement compared to model-based and data-driven baselines, respectively.",
keywords = "Compressive sensing, Feature fusion, Laplacian pyramid, Reconstruction, Reconstructive adversarial network",
author = "Kai Xu and Zhikang Zhang and Fengbo Ren",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-030-01249-6_30",
language = "English (US)",
isbn = "9783030012489",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "491--507",
editor = "Martial Hebert and Vittorio Ferrari and Cristian Sminchisescu and Yair Weiss",
booktitle = "Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings",

}

TY - GEN

T1 - LAPRAN

T2 - A scalable laplacian pyramid reconstructive adversarial network for flexible compressive sensing reconstruction

AU - Xu, Kai

AU - Zhang, Zhikang

AU - Ren, Fengbo

PY - 2018/1/1

Y1 - 2018/1/1

N2 - This paper addresses the single-image compressive sensing (CS) and reconstruction problem. We propose a scalable Laplacian pyramid reconstructive adversarial network (LAPRAN) that enables high-fidelity, flexible and fast CS images reconstruction. LAPRAN progressively reconstructs an image following the concept of the Laplacian pyramid through multiple stages of reconstructive adversarial networks (RANs). At each pyramid level, CS measurements are fused with a contextual latent vector to generate a high-frequency image residual. Consequently, LAPRAN can produce hierarchies of reconstructed images and each with an incremental resolution and improved quality. The scalable pyramid structure of LAPRAN enables high-fidelity CS reconstruction with a flexible resolution that is adaptive to a wide range of compression ratios (CRs), which is infeasible with existing methods. Experimental results on multiple public datasets show that LAPRAN offers an average 7.47 dB and 5.98 dB PSNR, and an average 57.93 % and 33.20 % SSIM improvement compared to model-based and data-driven baselines, respectively.

AB - This paper addresses the single-image compressive sensing (CS) and reconstruction problem. We propose a scalable Laplacian pyramid reconstructive adversarial network (LAPRAN) that enables high-fidelity, flexible and fast CS images reconstruction. LAPRAN progressively reconstructs an image following the concept of the Laplacian pyramid through multiple stages of reconstructive adversarial networks (RANs). At each pyramid level, CS measurements are fused with a contextual latent vector to generate a high-frequency image residual. Consequently, LAPRAN can produce hierarchies of reconstructed images and each with an incremental resolution and improved quality. The scalable pyramid structure of LAPRAN enables high-fidelity CS reconstruction with a flexible resolution that is adaptive to a wide range of compression ratios (CRs), which is infeasible with existing methods. Experimental results on multiple public datasets show that LAPRAN offers an average 7.47 dB and 5.98 dB PSNR, and an average 57.93 % and 33.20 % SSIM improvement compared to model-based and data-driven baselines, respectively.

KW - Compressive sensing

KW - Feature fusion

KW - Laplacian pyramid

KW - Reconstruction

KW - Reconstructive adversarial network

UR - http://www.scopus.com/inward/record.url?scp=85055104170&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055104170&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-01249-6_30

DO - 10.1007/978-3-030-01249-6_30

M3 - Conference contribution

SN - 9783030012489

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 491

EP - 507

BT - Computer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings

A2 - Hebert, Martial

A2 - Ferrari, Vittorio

A2 - Sminchisescu, Cristian

A2 - Weiss, Yair

PB - Springer Verlag

ER -