Image inpainting based on self-organizing maps by using multi-agent implementation

Margarita Favorskaya, Lakhmi JAIN, Andrey Bolgov

Research output: A Conference proceeding or a Chapter in BookConference contribution

5 Citations (Scopus)
2 Downloads (Pure)

Abstract

The image inpainting is a well-known task of visual editing. However, the efficiency strongly depends on sizes and textural neighborhood of "missing" area. Various methods of image inpainting exist, among which the Kohonen Self-Organizing Map (SOM) network as a mean of unsupervised learning is widely used. The weaknesses of the Kohonen SOM network such as the necessity for tuning of algorithm parameters and the low computational speed caused the application of multiagent system with a multi-mapping possibility and a parallel processing by the identical agents. During experiments, it was shown that the preliminary image segmentation and the creation of the SOMs for each type of homogeneous textures provide better results in comparison with the classical SOM application. Also the optimal number of inpainting agents was determined. The quality of inpainting was estimated by several metrics, and good results were obtained in complex images.

Original languageEnglish
Title of host publicationKnowledge-Based and Intelligent Information & Engineering Systems 18th Annual Conference, KES-2014 Gdynia, Poland, September 2014: Proceedings
EditorsPiotr Jedrzejowicz, Ireneusz Czarnowski, Robert J Howlett, Lakhmi C Jain
Place of PublicationNetherlands
PublisherElsevier
Pages861-870
Number of pages10
Volume35
DOIs
Publication statusPublished - 2014
Event18th International Conference on Knowledge-based and Intelligent Information and Engineering Systems - Gdinya, Gdinya, Poland
Duration: 15 Sep 201417 Sep 2014
http://kes2014.kesinternational.org/

Publication series

NameProcedia Computer Science
PublisherElsevier
Volume35
ISSN (Print)1877-0509

Conference

Conference18th International Conference on Knowledge-based and Intelligent Information and Engineering Systems
CountryPoland
CityGdinya
Period15/09/1417/09/14
Internet address

Fingerprint

Self organizing maps
Unsupervised learning
Multi agent systems
Image segmentation
Tuning
Textures
Processing
Experiments

Cite this

Favorskaya, M., JAIN, L., & Bolgov, A. (2014). Image inpainting based on self-organizing maps by using multi-agent implementation. In P. Jedrzejowicz, I. Czarnowski, R. J. Howlett, & L. C. Jain (Eds.), Knowledge-Based and Intelligent Information & Engineering Systems 18th Annual Conference, KES-2014 Gdynia, Poland, September 2014: Proceedings (Vol. 35, pp. 861-870). (Procedia Computer Science; Vol. 35). Netherlands: Elsevier. https://doi.org/10.1016/j.procs.2014.08.253
Favorskaya, Margarita ; JAIN, Lakhmi ; Bolgov, Andrey. / Image inpainting based on self-organizing maps by using multi-agent implementation. Knowledge-Based and Intelligent Information & Engineering Systems 18th Annual Conference, KES-2014 Gdynia, Poland, September 2014: Proceedings. editor / Piotr Jedrzejowicz ; Ireneusz Czarnowski ; Robert J Howlett ; Lakhmi C Jain. Vol. 35 Netherlands : Elsevier, 2014. pp. 861-870 (Procedia Computer Science).
@inproceedings{b98c72a3f73247c5ba08dd76bc4a9ea5,
title = "Image inpainting based on self-organizing maps by using multi-agent implementation",
abstract = "The image inpainting is a well-known task of visual editing. However, the efficiency strongly depends on sizes and textural neighborhood of {"}missing{"} area. Various methods of image inpainting exist, among which the Kohonen Self-Organizing Map (SOM) network as a mean of unsupervised learning is widely used. The weaknesses of the Kohonen SOM network such as the necessity for tuning of algorithm parameters and the low computational speed caused the application of multiagent system with a multi-mapping possibility and a parallel processing by the identical agents. During experiments, it was shown that the preliminary image segmentation and the creation of the SOMs for each type of homogeneous textures provide better results in comparison with the classical SOM application. Also the optimal number of inpainting agents was determined. The quality of inpainting was estimated by several metrics, and good results were obtained in complex images.",
keywords = "Kohonen SOM networks, clustering, image inpainting, texture, multi-agent system",
author = "Margarita Favorskaya and Lakhmi JAIN and Andrey Bolgov",
year = "2014",
doi = "10.1016/j.procs.2014.08.253",
language = "English",
volume = "35",
series = "Procedia Computer Science",
publisher = "Elsevier",
pages = "861--870",
editor = "Piotr Jedrzejowicz and Ireneusz Czarnowski and Howlett, {Robert J} and Jain, {Lakhmi C}",
booktitle = "Knowledge-Based and Intelligent Information & Engineering Systems 18th Annual Conference, KES-2014 Gdynia, Poland, September 2014: Proceedings",
address = "Netherlands",

}

Favorskaya, M, JAIN, L & Bolgov, A 2014, Image inpainting based on self-organizing maps by using multi-agent implementation. in P Jedrzejowicz, I Czarnowski, RJ Howlett & LC Jain (eds), Knowledge-Based and Intelligent Information & Engineering Systems 18th Annual Conference, KES-2014 Gdynia, Poland, September 2014: Proceedings. vol. 35, Procedia Computer Science, vol. 35, Elsevier, Netherlands, pp. 861-870, 18th International Conference on Knowledge-based and Intelligent Information and Engineering Systems, Gdinya, Poland, 15/09/14. https://doi.org/10.1016/j.procs.2014.08.253

Image inpainting based on self-organizing maps by using multi-agent implementation. / Favorskaya, Margarita; JAIN, Lakhmi; Bolgov, Andrey.

Knowledge-Based and Intelligent Information & Engineering Systems 18th Annual Conference, KES-2014 Gdynia, Poland, September 2014: Proceedings. ed. / Piotr Jedrzejowicz; Ireneusz Czarnowski; Robert J Howlett; Lakhmi C Jain. Vol. 35 Netherlands : Elsevier, 2014. p. 861-870 (Procedia Computer Science; Vol. 35).

Research output: A Conference proceeding or a Chapter in BookConference contribution

TY - GEN

T1 - Image inpainting based on self-organizing maps by using multi-agent implementation

AU - Favorskaya, Margarita

AU - JAIN, Lakhmi

AU - Bolgov, Andrey

PY - 2014

Y1 - 2014

N2 - The image inpainting is a well-known task of visual editing. However, the efficiency strongly depends on sizes and textural neighborhood of "missing" area. Various methods of image inpainting exist, among which the Kohonen Self-Organizing Map (SOM) network as a mean of unsupervised learning is widely used. The weaknesses of the Kohonen SOM network such as the necessity for tuning of algorithm parameters and the low computational speed caused the application of multiagent system with a multi-mapping possibility and a parallel processing by the identical agents. During experiments, it was shown that the preliminary image segmentation and the creation of the SOMs for each type of homogeneous textures provide better results in comparison with the classical SOM application. Also the optimal number of inpainting agents was determined. The quality of inpainting was estimated by several metrics, and good results were obtained in complex images.

AB - The image inpainting is a well-known task of visual editing. However, the efficiency strongly depends on sizes and textural neighborhood of "missing" area. Various methods of image inpainting exist, among which the Kohonen Self-Organizing Map (SOM) network as a mean of unsupervised learning is widely used. The weaknesses of the Kohonen SOM network such as the necessity for tuning of algorithm parameters and the low computational speed caused the application of multiagent system with a multi-mapping possibility and a parallel processing by the identical agents. During experiments, it was shown that the preliminary image segmentation and the creation of the SOMs for each type of homogeneous textures provide better results in comparison with the classical SOM application. Also the optimal number of inpainting agents was determined. The quality of inpainting was estimated by several metrics, and good results were obtained in complex images.

KW - Kohonen SOM networks

KW - clustering

KW - image inpainting

KW - texture

KW - multi-agent system

U2 - 10.1016/j.procs.2014.08.253

DO - 10.1016/j.procs.2014.08.253

M3 - Conference contribution

VL - 35

T3 - Procedia Computer Science

SP - 861

EP - 870

BT - Knowledge-Based and Intelligent Information & Engineering Systems 18th Annual Conference, KES-2014 Gdynia, Poland, September 2014: Proceedings

A2 - Jedrzejowicz, Piotr

A2 - Czarnowski, Ireneusz

A2 - Howlett, Robert J

A2 - Jain, Lakhmi C

PB - Elsevier

CY - Netherlands

ER -

Favorskaya M, JAIN L, Bolgov A. Image inpainting based on self-organizing maps by using multi-agent implementation. In Jedrzejowicz P, Czarnowski I, Howlett RJ, Jain LC, editors, Knowledge-Based and Intelligent Information & Engineering Systems 18th Annual Conference, KES-2014 Gdynia, Poland, September 2014: Proceedings. Vol. 35. Netherlands: Elsevier. 2014. p. 861-870. (Procedia Computer Science). https://doi.org/10.1016/j.procs.2014.08.253