Robust MMSE video decoding: Theory and practical implementations

Chang Su Kim; Jongwon Kim; Ioannis Katsavounidis; C. C. Jay Kuo

doi:10.1109/TCSVT.2004.839994

Robust MMSE video decoding: Theory and practical implementations

Chang Su Kim, Jongwon Kim, Ioannis Katsavounidis, C. C. Jay Kuo

Research output: Contribution to journal › Article › peer-review

9 Citations (Scopus)

Abstract

A novel video decoding algorithm based on the minimum mean square error (MMSE) criterion is investigated in this research. To alleviate the effect of transmission errors, we first develop an error propagation model to estimate and track the mean square error (MSE) of each pixel in the decoder. Then, the proposed video decoding algorithm adjusts the reconstruction of each pixel adaptively according to fluctuating channel conditions. More specifically, the decoder reconstructs a pixel in the kth frame F_k by using a weighted sum of two pixels in frames F_k-1 and F_k-2, respectively, where their weights are adaptively selected to minimize the MSE of the reconstructed pixel by using the error propagation model. Extensive simulation results performed on standard H.263 bit streams demonstrate that the MMSE-based concealment algorithm yields a better performance than the conventional method, even if the encoder transmits a single motion vector per block. Moreover, the proposed MMSE decoding algorithm significantly enhances the error resilient capability of the double-vector motion compensation (DMC) algorithm, where two motion vectors are sent per block.

Original language	English
Pages (from-to)	39-50
Number of pages	12
Journal	IEEE Transactions on Circuits and Systems for Video Technology
Volume	15
Issue number	1
DOIs	https://doi.org/10.1109/TCSVT.2004.839994
Publication status	Published - 2005 Jan
Externally published	Yes

Bibliographical note

Funding Information:
Manuscript received May 21, 2002; revised June 5, 2003. This work was supported in part by InterVideo, Inc. This paper was recommended by Associate Editor Q. Tian. C.-S. Kim is with the Department of Information Engineering, Chinese University of Hong Kong, Shatin, Hong Kong (e-mail: [email protected]). J. Kim is with the Department of Information and Communications, Kwangju Institute of Science and Technology, Kwangju 500-712, Korea (e-mail: [email protected]). I. Katsavounidis is with InterVideo Inc., Fremont, CA 94539 USA (e-mail: [email protected]). C.-C. J. Kuo is with the Department of Electrical Engineering, University of Southern California, Los Angeles, CA 90089-2564 USA (e-mail: [email protected]). Digital Object Identifier 10.1109/TCSVT.2004.839994

Keywords

Double-vector motion compensation
Error concealment
H.263
Minimum mean square error (MMSE) decoding
Robust video transmission

ASJC Scopus subject areas

Media Technology
Electrical and Electronic Engineering

Access to Document

10.1109/TCSVT.2004.839994

Cite this

@article{8039a7c1dd7b46cca3cf4dcad4445af9,

title = "Robust MMSE video decoding: Theory and practical implementations",

abstract = "A novel video decoding algorithm based on the minimum mean square error (MMSE) criterion is investigated in this research. To alleviate the effect of transmission errors, we first develop an error propagation model to estimate and track the mean square error (MSE) of each pixel in the decoder. Then, the proposed video decoding algorithm adjusts the reconstruction of each pixel adaptively according to fluctuating channel conditions. More specifically, the decoder reconstructs a pixel in the kth frame Fk by using a weighted sum of two pixels in frames Fk-1 and Fk-2, respectively, where their weights are adaptively selected to minimize the MSE of the reconstructed pixel by using the error propagation model. Extensive simulation results performed on standard H.263 bit streams demonstrate that the MMSE-based concealment algorithm yields a better performance than the conventional method, even if the encoder transmits a single motion vector per block. Moreover, the proposed MMSE decoding algorithm significantly enhances the error resilient capability of the double-vector motion compensation (DMC) algorithm, where two motion vectors are sent per block.",

keywords = "Double-vector motion compensation, Error concealment, H.263, Minimum mean square error (MMSE) decoding, Robust video transmission",

author = "Kim, {Chang Su} and Jongwon Kim and Ioannis Katsavounidis and {Jay Kuo}, {C. C.}",

note = "Funding Information: Manuscript received May 21, 2002; revised June 5, 2003. This work was supported in part by InterVideo, Inc. This paper was recommended by Associate Editor Q. Tian. C.-S. Kim is with the Department of Information Engineering, Chinese University of Hong Kong, Shatin, Hong Kong (e-mail: [email protected]). J. Kim is with the Department of Information and Communications, Kwangju Institute of Science and Technology, Kwangju 500-712, Korea (e-mail: [email protected]). I. Katsavounidis is with InterVideo Inc., Fremont, CA 94539 USA (e-mail: [email protected]). C.-C. J. Kuo is with the Department of Electrical Engineering, University of Southern California, Los Angeles, CA 90089-2564 USA (e-mail: [email protected]). Digital Object Identifier 10.1109/TCSVT.2004.839994",

year = "2005",

month = jan,

doi = "10.1109/TCSVT.2004.839994",

language = "English",

volume = "15",

pages = "39--50",

journal = "IEEE Transactions on Circuits and Systems for Video Technology",

issn = "1051-8215",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "1",

}

TY - JOUR

T1 - Robust MMSE video decoding

T2 - Theory and practical implementations

AU - Kim, Chang Su

AU - Kim, Jongwon

AU - Katsavounidis, Ioannis

AU - Jay Kuo, C. C.

N1 - Funding Information: Manuscript received May 21, 2002; revised June 5, 2003. This work was supported in part by InterVideo, Inc. This paper was recommended by Associate Editor Q. Tian. C.-S. Kim is with the Department of Information Engineering, Chinese University of Hong Kong, Shatin, Hong Kong (e-mail: [email protected]). J. Kim is with the Department of Information and Communications, Kwangju Institute of Science and Technology, Kwangju 500-712, Korea (e-mail: [email protected]). I. Katsavounidis is with InterVideo Inc., Fremont, CA 94539 USA (e-mail: [email protected]). C.-C. J. Kuo is with the Department of Electrical Engineering, University of Southern California, Los Angeles, CA 90089-2564 USA (e-mail: [email protected]). Digital Object Identifier 10.1109/TCSVT.2004.839994

PY - 2005/1

Y1 - 2005/1

N2 - A novel video decoding algorithm based on the minimum mean square error (MMSE) criterion is investigated in this research. To alleviate the effect of transmission errors, we first develop an error propagation model to estimate and track the mean square error (MSE) of each pixel in the decoder. Then, the proposed video decoding algorithm adjusts the reconstruction of each pixel adaptively according to fluctuating channel conditions. More specifically, the decoder reconstructs a pixel in the kth frame Fk by using a weighted sum of two pixels in frames Fk-1 and Fk-2, respectively, where their weights are adaptively selected to minimize the MSE of the reconstructed pixel by using the error propagation model. Extensive simulation results performed on standard H.263 bit streams demonstrate that the MMSE-based concealment algorithm yields a better performance than the conventional method, even if the encoder transmits a single motion vector per block. Moreover, the proposed MMSE decoding algorithm significantly enhances the error resilient capability of the double-vector motion compensation (DMC) algorithm, where two motion vectors are sent per block.

AB - A novel video decoding algorithm based on the minimum mean square error (MMSE) criterion is investigated in this research. To alleviate the effect of transmission errors, we first develop an error propagation model to estimate and track the mean square error (MSE) of each pixel in the decoder. Then, the proposed video decoding algorithm adjusts the reconstruction of each pixel adaptively according to fluctuating channel conditions. More specifically, the decoder reconstructs a pixel in the kth frame Fk by using a weighted sum of two pixels in frames Fk-1 and Fk-2, respectively, where their weights are adaptively selected to minimize the MSE of the reconstructed pixel by using the error propagation model. Extensive simulation results performed on standard H.263 bit streams demonstrate that the MMSE-based concealment algorithm yields a better performance than the conventional method, even if the encoder transmits a single motion vector per block. Moreover, the proposed MMSE decoding algorithm significantly enhances the error resilient capability of the double-vector motion compensation (DMC) algorithm, where two motion vectors are sent per block.

KW - Double-vector motion compensation

KW - Error concealment

KW - H.263

KW - Minimum mean square error (MMSE) decoding

KW - Robust video transmission

UR - http://www.scopus.com/inward/record.url?scp=12344329319&partnerID=8YFLogxK

U2 - 10.1109/TCSVT.2004.839994

DO - 10.1109/TCSVT.2004.839994

M3 - Article

AN - SCOPUS:12344329319

SN - 1051-8215

VL - 15

SP - 39

EP - 50

JO - IEEE Transactions on Circuits and Systems for Video Technology

JF - IEEE Transactions on Circuits and Systems for Video Technology

IS - 1

ER -

Robust MMSE video decoding: Theory and practical implementations

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this