Rotation-invariant categorization of colour images using the Radon transform

Andrew Peter Paplinski

Rotation-invariant categorization of colour images using the Radon transform

Andrew Peter Paplinski

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

Abstract

We have derived a novel rotation-invariant feature vector, or signature, for two-variable functions like images. The feature vector is calculated as an angular integral of the Radon transform of the function. Three such feature vectors are calculated for each colour image. Subsequently, these feature vectors are used to categorize colour images in a network of 3+1 self-organizing modules. The 3-D ‘labels’ produced by the first level modules are used by the second level modules and can be thought of as a “universal neuronal code”. The network is trained for un-rotated images and then tested for rotated images. It has been demonstrated that rotation of images by the angles included in calculation of the Radon transform results
in the perfect categorization. For the angles in between, that is, those not included in the Radon transform, a small shift in
categorization might occur, keeping, however, the objects well inside their clusters. Since calculation of the rotation-invariant feature vectors is very simple and involves only summations of signals (pixel values), hence very fast, it is postulated that such
a mechanism might be included in the biological vision systems.

Original language	English
Title of host publication	2012 International Joint Conference on Neural Networks
Editors	Kate Smith-Miles
Place of Publication	Piscataway NJ USA
Publisher	IEEE, Institute of Electrical and Electronics Engineers
Pages	1408 - 1413
Number of pages	6
ISBN (Print)	9781467314909
Publication status	Published - 2012
Event	IEEE International Joint Conference on Neural Networks 2012 - Brisbane Convention & Exhibition Centre, Brisbane, Australia Duration: 10 Jun 2012 → 15 Jun 2012 https://ieeexplore.ieee.org/xpl/conhome/6241467/proceeding (Proceedings)

Conference

Conference	IEEE International Joint Conference on Neural Networks 2012
Abbreviated title	IJCNN 2012
Country/Territory	Australia
City	Brisbane
Period	10/06/12 → 15/06/12
Internet address	https://ieeexplore.ieee.org/xpl/conhome/6241467/proceeding (Proceedings)

Cite this

@inproceedings{89bb433bf6224bb097374f3feeb83127,

title = "Rotation-invariant categorization of colour images using the Radon transform",

abstract = "We have derived a novel rotation-invariant feature vector, or signature, for two-variable functions like images. The feature vector is calculated as an angular integral of the Radon transform of the function. Three such feature vectors are calculated for each colour image. Subsequently, these feature vectors are used to categorize colour images in a network of 3+1 self-organizing modules. The 3-D {\textquoteleft}labels{\textquoteright} produced by the first level modules are used by the second level modules and can be thought of as a “universal neuronal code”. The network is trained for un-rotated images and then tested for rotated images. It has been demonstrated that rotation of images by the angles included in calculation of the Radon transform resultsin the perfect categorization. For the angles in between, that is, those not included in the Radon transform, a small shift incategorization might occur, keeping, however, the objects well inside their clusters. Since calculation of the rotation-invariant feature vectors is very simple and involves only summations of signals (pixel values), hence very fast, it is postulated that sucha mechanism might be included in the biological vision systems.",

author = "Paplinski, {Andrew Peter}",

year = "2012",

language = "English",

isbn = "9781467314909",

pages = "1408 -- 1413",

editor = "Kate Smith-Miles",

booktitle = "2012 International Joint Conference on Neural Networks",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

address = "United States of America",

note = "IEEE International Joint Conference on Neural Networks 2012, IJCNN 2012 ; Conference date: 10-06-2012 Through 15-06-2012",

url = "https://ieeexplore.ieee.org/xpl/conhome/6241467/proceeding",

}

Paplinski, AP 2012, Rotation-invariant categorization of colour images using the Radon transform. in K Smith-Miles (ed.), 2012 International Joint Conference on Neural Networks. IEEE, Institute of Electrical and Electronics Engineers, Piscataway NJ USA, pp. 1408 - 1413, IEEE International Joint Conference on Neural Networks 2012, Brisbane, Queensland, Australia, 10/06/12.

Rotation-invariant categorization of colour images using the Radon transform. / Paplinski, Andrew Peter.
2012 International Joint Conference on Neural Networks. ed. / Kate Smith-Miles. Piscataway NJ USA: IEEE, Institute of Electrical and Electronics Engineers, 2012. p. 1408 - 1413.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - Rotation-invariant categorization of colour images using the Radon transform

AU - Paplinski, Andrew Peter

PY - 2012

Y1 - 2012

N2 - We have derived a novel rotation-invariant feature vector, or signature, for two-variable functions like images. The feature vector is calculated as an angular integral of the Radon transform of the function. Three such feature vectors are calculated for each colour image. Subsequently, these feature vectors are used to categorize colour images in a network of 3+1 self-organizing modules. The 3-D ‘labels’ produced by the first level modules are used by the second level modules and can be thought of as a “universal neuronal code”. The network is trained for un-rotated images and then tested for rotated images. It has been demonstrated that rotation of images by the angles included in calculation of the Radon transform resultsin the perfect categorization. For the angles in between, that is, those not included in the Radon transform, a small shift incategorization might occur, keeping, however, the objects well inside their clusters. Since calculation of the rotation-invariant feature vectors is very simple and involves only summations of signals (pixel values), hence very fast, it is postulated that sucha mechanism might be included in the biological vision systems.

AB - We have derived a novel rotation-invariant feature vector, or signature, for two-variable functions like images. The feature vector is calculated as an angular integral of the Radon transform of the function. Three such feature vectors are calculated for each colour image. Subsequently, these feature vectors are used to categorize colour images in a network of 3+1 self-organizing modules. The 3-D ‘labels’ produced by the first level modules are used by the second level modules and can be thought of as a “universal neuronal code”. The network is trained for un-rotated images and then tested for rotated images. It has been demonstrated that rotation of images by the angles included in calculation of the Radon transform resultsin the perfect categorization. For the angles in between, that is, those not included in the Radon transform, a small shift incategorization might occur, keeping, however, the objects well inside their clusters. Since calculation of the rotation-invariant feature vectors is very simple and involves only summations of signals (pixel values), hence very fast, it is postulated that sucha mechanism might be included in the biological vision systems.

M3 - Conference Paper

SN - 9781467314909

SP - 1408

EP - 1413

BT - 2012 International Joint Conference on Neural Networks

A2 - Smith-Miles, Kate

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Piscataway NJ USA

T2 - IEEE International Joint Conference on Neural Networks 2012

Y2 - 10 June 2012 through 15 June 2012

ER -