Compressing moving pictures using the APEX neural principal component extractor

K. I. Diamantaras; S. Y. Kung

doi:10.1109/NNSP.1993.471856

Compressing moving pictures using the APEX neural principal component extractor

K. I. Diamantaras, S. Y. Kung

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

In this paper we propose an application of the optimal Karhunen-Lokve Transform in place of the traditional DCT for compressing Intra-frames in the MPEG protocol. The I-frames attain the smallest compression ratio since they are coded without reference to any other frames. The difficulty of KLT that additional bit-rate is required to make the image-dependent transform basis known to the decoder is overcome by using the KLT basis of the previous I-or P-frame which we argue, is very similar to the basis of the current frame. The previous frame is already known to the decoder, therefore no additional information needs to be sent out. Paying attention to the non-stationary statistics of images we also propose to split the I-frames in N parts and use a dedicated transform basis for each part. Since the KLT basis must be updated continuously we make use of the adaptive Principal Component extractor network (APEX) that can incrementally estimate the new basis for the next frame. Our simulations for two sports sequences indicate that a 10% to 30% error improvement is achieved using the same number of components for DCT and KLT. Furthermore, the MPEG bit-stream format does not have to change to accommodate the proposed technique.

Original language	English (US)
Title of host publication	Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993
Editors	C.A. Kamm, G.M. Kuhn, R. Chellappa, B. Yoon, S.Y. Kung
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	321-330
Number of pages	10
ISBN (Electronic)	0780309286, 9780780309289
DOIs	https://doi.org/10.1109/NNSP.1993.471856
State	Published - 1993
Event	1993 3rd IEEE-SP Workshop on Neural Networks for Signal Processing, NNSP 1993 - Linthicum Heights, United States Duration: Sep 6 1993 → Sep 9 1993

Publication series

Name	Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993

Conference

Conference	1993 3rd IEEE-SP Workshop on Neural Networks for Signal Processing, NNSP 1993
Country/Territory	United States
City	Linthicum Heights
Period	9/6/93 → 9/9/93

All Science Journal Classification (ASJC) codes

Artificial Intelligence
Signal Processing
Safety, Risk, Reliability and Quality

Access to Document

10.1109/NNSP.1993.471856

Cite this

Diamantaras, K. I., & Kung, S. Y. (1993). Compressing moving pictures using the APEX neural principal component extractor. In C. A. Kamm, G. M. Kuhn, R. Chellappa, B. Yoon, & S. Y. Kung (Eds.), Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993 (pp. 321-330). Article 471856 (Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/NNSP.1993.471856

Diamantaras, K. I. ; Kung, S. Y. / Compressing moving pictures using the APEX neural principal component extractor. Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993. editor / C.A. Kamm ; G.M. Kuhn ; R. Chellappa ; B. Yoon ; S.Y. Kung. Institute of Electrical and Electronics Engineers Inc., 1993. pp. 321-330 (Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993).

@inproceedings{8aa1c0e41ee94d5b8f1c443554b7ac5a,

title = "Compressing moving pictures using the APEX neural principal component extractor",

abstract = "In this paper we propose an application of the optimal Karhunen-Lokve Transform in place of the traditional DCT for compressing Intra-frames in the MPEG protocol. The I-frames attain the smallest compression ratio since they are coded without reference to any other frames. The difficulty of KLT that additional bit-rate is required to make the image-dependent transform basis known to the decoder is overcome by using the KLT basis of the previous I-or P-frame which we argue, is very similar to the basis of the current frame. The previous frame is already known to the decoder, therefore no additional information needs to be sent out. Paying attention to the non-stationary statistics of images we also propose to split the I-frames in N parts and use a dedicated transform basis for each part. Since the KLT basis must be updated continuously we make use of the adaptive Principal Component extractor network (APEX) that can incrementally estimate the new basis for the next frame. Our simulations for two sports sequences indicate that a 10% to 30% error improvement is achieved using the same number of components for DCT and KLT. Furthermore, the MPEG bit-stream format does not have to change to accommodate the proposed technique.",

author = "Diamantaras, {K. I.} and Kung, {S. Y.}",

note = "Publisher Copyright: {\textcopyright} 1993 IEEE.; 1993 3rd IEEE-SP Workshop on Neural Networks for Signal Processing, NNSP 1993 ; Conference date: 06-09-1993 Through 09-09-1993",

year = "1993",

doi = "10.1109/NNSP.1993.471856",

language = "English (US)",

series = "Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993",

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

pages = "321--330",

editor = "C.A. Kamm and G.M. Kuhn and R. Chellappa and B. Yoon and S.Y. Kung",

booktitle = "Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993",

address = "United States",

}

Diamantaras, KI & Kung, SY 1993, Compressing moving pictures using the APEX neural principal component extractor. in CA Kamm, GM Kuhn, R Chellappa, B Yoon & SY Kung (eds), Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993., 471856, Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993, Institute of Electrical and Electronics Engineers Inc., pp. 321-330, 1993 3rd IEEE-SP Workshop on Neural Networks for Signal Processing, NNSP 1993, Linthicum Heights, United States, 9/6/93. https://doi.org/10.1109/NNSP.1993.471856

Compressing moving pictures using the APEX neural principal component extractor. / Diamantaras, K. I.; Kung, S. Y.
Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993. ed. / C.A. Kamm; G.M. Kuhn; R. Chellappa; B. Yoon; S.Y. Kung. Institute of Electrical and Electronics Engineers Inc., 1993. p. 321-330 471856 (Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Compressing moving pictures using the APEX neural principal component extractor

AU - Diamantaras, K. I.

AU - Kung, S. Y.

PY - 1993

Y1 - 1993

N2 - In this paper we propose an application of the optimal Karhunen-Lokve Transform in place of the traditional DCT for compressing Intra-frames in the MPEG protocol. The I-frames attain the smallest compression ratio since they are coded without reference to any other frames. The difficulty of KLT that additional bit-rate is required to make the image-dependent transform basis known to the decoder is overcome by using the KLT basis of the previous I-or P-frame which we argue, is very similar to the basis of the current frame. The previous frame is already known to the decoder, therefore no additional information needs to be sent out. Paying attention to the non-stationary statistics of images we also propose to split the I-frames in N parts and use a dedicated transform basis for each part. Since the KLT basis must be updated continuously we make use of the adaptive Principal Component extractor network (APEX) that can incrementally estimate the new basis for the next frame. Our simulations for two sports sequences indicate that a 10% to 30% error improvement is achieved using the same number of components for DCT and KLT. Furthermore, the MPEG bit-stream format does not have to change to accommodate the proposed technique.

AB - In this paper we propose an application of the optimal Karhunen-Lokve Transform in place of the traditional DCT for compressing Intra-frames in the MPEG protocol. The I-frames attain the smallest compression ratio since they are coded without reference to any other frames. The difficulty of KLT that additional bit-rate is required to make the image-dependent transform basis known to the decoder is overcome by using the KLT basis of the previous I-or P-frame which we argue, is very similar to the basis of the current frame. The previous frame is already known to the decoder, therefore no additional information needs to be sent out. Paying attention to the non-stationary statistics of images we also propose to split the I-frames in N parts and use a dedicated transform basis for each part. Since the KLT basis must be updated continuously we make use of the adaptive Principal Component extractor network (APEX) that can incrementally estimate the new basis for the next frame. Our simulations for two sports sequences indicate that a 10% to 30% error improvement is achieved using the same number of components for DCT and KLT. Furthermore, the MPEG bit-stream format does not have to change to accommodate the proposed technique.

UR - http://www.scopus.com/inward/record.url?scp=85064822026&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064822026&partnerID=8YFLogxK

U2 - 10.1109/NNSP.1993.471856

DO - 10.1109/NNSP.1993.471856

M3 - Conference contribution

AN - SCOPUS:85064822026

T3 - Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993

SP - 321

EP - 330

BT - Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993

A2 - Kamm, C.A.

A2 - Kuhn, G.M.

A2 - Chellappa, R.

A2 - Yoon, B.

A2 - Kung, S.Y.

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 1993 3rd IEEE-SP Workshop on Neural Networks for Signal Processing, NNSP 1993

Y2 - 6 September 1993 through 9 September 1993

ER -

Diamantaras KI, Kung SY. Compressing moving pictures using the APEX neural principal component extractor. In Kamm CA, Kuhn GM, Chellappa R, Yoon B, Kung SY, editors, Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993. Institute of Electrical and Electronics Engineers Inc. 1993. p. 321-330. 471856. (Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE Workshop, NNSP 1993). doi: 10.1109/NNSP.1993.471856

Compressing moving pictures using the APEX neural principal component extractor

Abstract

Publication series

Conference

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this