Adaptive encoding of a videoconference image sequence via neural networks

C. N. Manikopoulos, George Antoniou

Research output: Contribution to journalArticle

Abstract

A new method for encoding a videoconference image sequence, termed adaptive neural net vector quantisation (ANNVQ), has been derived. It is based on Kohonen's self-organised feature maps, a neural network type clustering algorithm. The new method differs from it, in that after training the initial codebook, a modified form of adaptation resumes, in order to respond to scene changes and motion. The main advantages are high image quality with modest bit rates and effective adaptation to motion and scene changes, with the capability to quickly adjust the instantaneous bit rate in order to keep the image quality constant. This is a good match to packet switched networks where variable bit rate and uniform image quality are highly desirable. Simulation experiments have been carried out with 4 × 4 blocks of pixels from an image sequence consisting of 20 frames of size 112 × 96 pixels each. With a codebook size of 512, ANNVQ results in high image quality upon image reconstruction, with peak signal-to-noise ratio (PSNR) of about 36 to 37 dB, at coding bit rates of about 0.50 bit/pixel. This compares quite favourably with classical vector quantisation at a similar bit rate. Moreover, this value of PSNR remains approximately constant, even when encoding image frames with considerable motion.

Original languageEnglish
Pages (from-to)233-241
Number of pages9
JournalJournal of Electrical and Electronics Engineering, Australia
Volume12
Issue number3
StatePublished - 1 Sep 1992

Fingerprint

Neural networks
Image quality
Vector quantization
Pixels
Signal to noise ratio
Packet networks
Image reconstruction
Clustering algorithms
Experiments

Cite this

@article{ba7b1af5a90143999810903996b4b9e1,
title = "Adaptive encoding of a videoconference image sequence via neural networks",
abstract = "A new method for encoding a videoconference image sequence, termed adaptive neural net vector quantisation (ANNVQ), has been derived. It is based on Kohonen's self-organised feature maps, a neural network type clustering algorithm. The new method differs from it, in that after training the initial codebook, a modified form of adaptation resumes, in order to respond to scene changes and motion. The main advantages are high image quality with modest bit rates and effective adaptation to motion and scene changes, with the capability to quickly adjust the instantaneous bit rate in order to keep the image quality constant. This is a good match to packet switched networks where variable bit rate and uniform image quality are highly desirable. Simulation experiments have been carried out with 4 × 4 blocks of pixels from an image sequence consisting of 20 frames of size 112 × 96 pixels each. With a codebook size of 512, ANNVQ results in high image quality upon image reconstruction, with peak signal-to-noise ratio (PSNR) of about 36 to 37 dB, at coding bit rates of about 0.50 bit/pixel. This compares quite favourably with classical vector quantisation at a similar bit rate. Moreover, this value of PSNR remains approximately constant, even when encoding image frames with considerable motion.",
author = "Manikopoulos, {C. N.} and George Antoniou",
year = "1992",
month = "9",
day = "1",
language = "English",
volume = "12",
pages = "233--241",
journal = "Journal of Electrical and Electronics Engineering, Australia",
issn = "0725-2986",
publisher = "Institution of Engineers (Australia)",
number = "3",

}

Adaptive encoding of a videoconference image sequence via neural networks. / Manikopoulos, C. N.; Antoniou, George.

In: Journal of Electrical and Electronics Engineering, Australia, Vol. 12, No. 3, 01.09.1992, p. 233-241.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Adaptive encoding of a videoconference image sequence via neural networks

AU - Manikopoulos, C. N.

AU - Antoniou, George

PY - 1992/9/1

Y1 - 1992/9/1

N2 - A new method for encoding a videoconference image sequence, termed adaptive neural net vector quantisation (ANNVQ), has been derived. It is based on Kohonen's self-organised feature maps, a neural network type clustering algorithm. The new method differs from it, in that after training the initial codebook, a modified form of adaptation resumes, in order to respond to scene changes and motion. The main advantages are high image quality with modest bit rates and effective adaptation to motion and scene changes, with the capability to quickly adjust the instantaneous bit rate in order to keep the image quality constant. This is a good match to packet switched networks where variable bit rate and uniform image quality are highly desirable. Simulation experiments have been carried out with 4 × 4 blocks of pixels from an image sequence consisting of 20 frames of size 112 × 96 pixels each. With a codebook size of 512, ANNVQ results in high image quality upon image reconstruction, with peak signal-to-noise ratio (PSNR) of about 36 to 37 dB, at coding bit rates of about 0.50 bit/pixel. This compares quite favourably with classical vector quantisation at a similar bit rate. Moreover, this value of PSNR remains approximately constant, even when encoding image frames with considerable motion.

AB - A new method for encoding a videoconference image sequence, termed adaptive neural net vector quantisation (ANNVQ), has been derived. It is based on Kohonen's self-organised feature maps, a neural network type clustering algorithm. The new method differs from it, in that after training the initial codebook, a modified form of adaptation resumes, in order to respond to scene changes and motion. The main advantages are high image quality with modest bit rates and effective adaptation to motion and scene changes, with the capability to quickly adjust the instantaneous bit rate in order to keep the image quality constant. This is a good match to packet switched networks where variable bit rate and uniform image quality are highly desirable. Simulation experiments have been carried out with 4 × 4 blocks of pixels from an image sequence consisting of 20 frames of size 112 × 96 pixels each. With a codebook size of 512, ANNVQ results in high image quality upon image reconstruction, with peak signal-to-noise ratio (PSNR) of about 36 to 37 dB, at coding bit rates of about 0.50 bit/pixel. This compares quite favourably with classical vector quantisation at a similar bit rate. Moreover, this value of PSNR remains approximately constant, even when encoding image frames with considerable motion.

UR - http://www.scopus.com/inward/record.url?scp=0026913127&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0026913127

VL - 12

SP - 233

EP - 241

JO - Journal of Electrical and Electronics Engineering, Australia

JF - Journal of Electrical and Electronics Engineering, Australia

SN - 0725-2986

IS - 3

ER -