Learned Neural Iterative Decoding for Lossy Image Compression Systems

Alexander G. Ororbia, Ankur Mali, Jian Wu, Scott O'Connell, William Dreese, David Jonathan Miller, C. Lee Giles

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

1 Citation (Scopus)

Abstract

For lossy image compression systems, we develop an algorithm, iterative refinement, to improve the decoder's reconstruction compared to standard decoding techniques. Specifically, we propose a recurrent neural network approach for nonlinear, iterative decoding. Our decoder, which works with any encoder, employs self-connected memory units that make use of causal and non-causal spatial context information to progressively reduce reconstruction error over a fixed number of steps. We experiment with variants of our estimator and find that iterative refinement consistently creates lower distortion images of higher perceptual quality compared to other approaches. Specifically, on the Kodak Lossless True Color Image Suite, we observe as much as a 0.871 decibel (dB) gain over JPEG, a 1.095 dB gain over JPEG 2000, and a 0.971 dB gain over a competitive neural model.

Original languageEnglish (US)
Title of host publicationProceedings - DCC 2019
Subtitle of host publication2019 Data Compression Conference
EditorsMichael W. Marcellin, Ali Bilgin, James A. Storer, Joan Serra-Sagrista
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3-12
Number of pages10
ISBN (Electronic)9781728106571
DOIs
StatePublished - May 10 2019
Event2019 Data Compression Conference, DCC 2019 - Snowbird, United States
Duration: Mar 26 2019Mar 29 2019

Publication series

NameData Compression Conference Proceedings
Volume2019-March
ISSN (Print)1068-0314

Conference

Conference2019 Data Compression Conference, DCC 2019
CountryUnited States
CitySnowbird
Period3/26/193/29/19

Fingerprint

Iterative decoding
Recurrent neural networks
Image compression
Decoding
Color
Data storage equipment
Experiments

All Science Journal Classification (ASJC) codes

  • Computer Networks and Communications

Cite this

Ororbia, A. G., Mali, A., Wu, J., O'Connell, S., Dreese, W., Miller, D. J., & Giles, C. L. (2019). Learned Neural Iterative Decoding for Lossy Image Compression Systems. In M. W. Marcellin, A. Bilgin, J. A. Storer, & J. Serra-Sagrista (Eds.), Proceedings - DCC 2019: 2019 Data Compression Conference (pp. 3-12). [8712688] (Data Compression Conference Proceedings; Vol. 2019-March). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/DCC.2019.00008
Ororbia, Alexander G. ; Mali, Ankur ; Wu, Jian ; O'Connell, Scott ; Dreese, William ; Miller, David Jonathan ; Giles, C. Lee. / Learned Neural Iterative Decoding for Lossy Image Compression Systems. Proceedings - DCC 2019: 2019 Data Compression Conference. editor / Michael W. Marcellin ; Ali Bilgin ; James A. Storer ; Joan Serra-Sagrista. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 3-12 (Data Compression Conference Proceedings).
@inproceedings{f7198400b10246bea9119232c7ee6ae1,
title = "Learned Neural Iterative Decoding for Lossy Image Compression Systems",
abstract = "For lossy image compression systems, we develop an algorithm, iterative refinement, to improve the decoder's reconstruction compared to standard decoding techniques. Specifically, we propose a recurrent neural network approach for nonlinear, iterative decoding. Our decoder, which works with any encoder, employs self-connected memory units that make use of causal and non-causal spatial context information to progressively reduce reconstruction error over a fixed number of steps. We experiment with variants of our estimator and find that iterative refinement consistently creates lower distortion images of higher perceptual quality compared to other approaches. Specifically, on the Kodak Lossless True Color Image Suite, we observe as much as a 0.871 decibel (dB) gain over JPEG, a 1.095 dB gain over JPEG 2000, and a 0.971 dB gain over a competitive neural model.",
author = "Ororbia, {Alexander G.} and Ankur Mali and Jian Wu and Scott O'Connell and William Dreese and Miller, {David Jonathan} and Giles, {C. Lee}",
year = "2019",
month = "5",
day = "10",
doi = "10.1109/DCC.2019.00008",
language = "English (US)",
series = "Data Compression Conference Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "3--12",
editor = "Marcellin, {Michael W.} and Ali Bilgin and Storer, {James A.} and Joan Serra-Sagrista",
booktitle = "Proceedings - DCC 2019",
address = "United States",

}

Ororbia, AG, Mali, A, Wu, J, O'Connell, S, Dreese, W, Miller, DJ & Giles, CL 2019, Learned Neural Iterative Decoding for Lossy Image Compression Systems. in MW Marcellin, A Bilgin, JA Storer & J Serra-Sagrista (eds), Proceedings - DCC 2019: 2019 Data Compression Conference., 8712688, Data Compression Conference Proceedings, vol. 2019-March, Institute of Electrical and Electronics Engineers Inc., pp. 3-12, 2019 Data Compression Conference, DCC 2019, Snowbird, United States, 3/26/19. https://doi.org/10.1109/DCC.2019.00008

Learned Neural Iterative Decoding for Lossy Image Compression Systems. / Ororbia, Alexander G.; Mali, Ankur; Wu, Jian; O'Connell, Scott; Dreese, William; Miller, David Jonathan; Giles, C. Lee.

Proceedings - DCC 2019: 2019 Data Compression Conference. ed. / Michael W. Marcellin; Ali Bilgin; James A. Storer; Joan Serra-Sagrista. Institute of Electrical and Electronics Engineers Inc., 2019. p. 3-12 8712688 (Data Compression Conference Proceedings; Vol. 2019-March).

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

TY - GEN

T1 - Learned Neural Iterative Decoding for Lossy Image Compression Systems

AU - Ororbia, Alexander G.

AU - Mali, Ankur

AU - Wu, Jian

AU - O'Connell, Scott

AU - Dreese, William

AU - Miller, David Jonathan

AU - Giles, C. Lee

PY - 2019/5/10

Y1 - 2019/5/10

N2 - For lossy image compression systems, we develop an algorithm, iterative refinement, to improve the decoder's reconstruction compared to standard decoding techniques. Specifically, we propose a recurrent neural network approach for nonlinear, iterative decoding. Our decoder, which works with any encoder, employs self-connected memory units that make use of causal and non-causal spatial context information to progressively reduce reconstruction error over a fixed number of steps. We experiment with variants of our estimator and find that iterative refinement consistently creates lower distortion images of higher perceptual quality compared to other approaches. Specifically, on the Kodak Lossless True Color Image Suite, we observe as much as a 0.871 decibel (dB) gain over JPEG, a 1.095 dB gain over JPEG 2000, and a 0.971 dB gain over a competitive neural model.

AB - For lossy image compression systems, we develop an algorithm, iterative refinement, to improve the decoder's reconstruction compared to standard decoding techniques. Specifically, we propose a recurrent neural network approach for nonlinear, iterative decoding. Our decoder, which works with any encoder, employs self-connected memory units that make use of causal and non-causal spatial context information to progressively reduce reconstruction error over a fixed number of steps. We experiment with variants of our estimator and find that iterative refinement consistently creates lower distortion images of higher perceptual quality compared to other approaches. Specifically, on the Kodak Lossless True Color Image Suite, we observe as much as a 0.871 decibel (dB) gain over JPEG, a 1.095 dB gain over JPEG 2000, and a 0.971 dB gain over a competitive neural model.

UR - http://www.scopus.com/inward/record.url?scp=85066316254&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066316254&partnerID=8YFLogxK

U2 - 10.1109/DCC.2019.00008

DO - 10.1109/DCC.2019.00008

M3 - Conference contribution

AN - SCOPUS:85066316254

T3 - Data Compression Conference Proceedings

SP - 3

EP - 12

BT - Proceedings - DCC 2019

A2 - Marcellin, Michael W.

A2 - Bilgin, Ali

A2 - Storer, James A.

A2 - Serra-Sagrista, Joan

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Ororbia AG, Mali A, Wu J, O'Connell S, Dreese W, Miller DJ et al. Learned Neural Iterative Decoding for Lossy Image Compression Systems. In Marcellin MW, Bilgin A, Storer JA, Serra-Sagrista J, editors, Proceedings - DCC 2019: 2019 Data Compression Conference. Institute of Electrical and Electronics Engineers Inc. 2019. p. 3-12. 8712688. (Data Compression Conference Proceedings). https://doi.org/10.1109/DCC.2019.00008