A Dipolar Competitive Neural Network for Video Segmentation

 

Abstract

This paper present a video segmentation method which separate pixels corresponding to foreground from those corresponding to background. The proposed background model consists of a competitive neural network based on dipoles, which is used to classify the pixels as background or foreground. Using this kind of neural networks permits an easy hardware implementation to achieve a real time processing with good results. The dipolar representation is designed to deal with the problem of estimating the directionality of data. Experimental results are provided by using the standard PETS dataset and compared with the mixture of Gaussians and background subtraction methods. © 2008 Springer-Verlag.

Citation

Please, cite this work as:

[Luq+08] R. M. Luque, D. López-Rodríguez, E. Domínguez, et al. “A Dipolar Competitive Neural Network for Video Segmentation”. In: Advances in Artificial Intelligence - IBERAMIA 2008, 11th Ibero-American Conference on AI, Lisbon, Portugal, October 14-17, 2008. Proceedings. Ed. by H. Geffner, R. Prada, I. M. Alexandre and N. David. Vol. 5290 LNAI. Lecture Notes in Computer Science. cited By 5; Conference of 11th Ibero-American Conference on Artificial Intelligence, IBERAMIA 2008 ; Conference Date: 14 October 2008 Through 17 October 2008; Conference Code:77561. Lisbon: Springer, 2008, pp. 103-112. DOI: 10.1007/978-3-540-88309-8_11. URL: https://doi.org/10.1007/978-3-540-88309-8_11.

BibTeX

@InProceedings{Luque2008,
     author = {Rafael Marcos Luque and Domingo López-Rodríguez and Enrique Domínguez and Esteban J. Palomo},
     booktitle = {Advances in Artificial Intelligence - {IBERAMIA} 2008, 11th Ibero-American Conference on AI, Lisbon, Portugal, October 14-17, 2008. Proceedings},
     title = {A Dipolar Competitive Neural Network for Video Segmentation},
     year = {2008},
     address = {Lisbon},
     editor = {Hector Geffner and Rui Prada and Isabel Machado Alexandre and Nuno David},
     note = {cited By 5; Conference of 11th Ibero-American Conference on Artificial Intelligence, IBERAMIA 2008 ; Conference Date: 14 October 2008 Through 17 October 2008; Conference Code:77561},
     pages = {103-112},
     publisher = {Springer},
     series = {Lecture Notes in Computer Science},
     volume = {5290 LNAI},
     abstract = {This paper present a video segmentation method which separate pixels corresponding to foreground from those corresponding to background. The proposed background model consists of a competitive neural network based on dipoles, which is used to classify the pixels as background or foreground. Using this kind of neural networks permits an easy hardware implementation to achieve a real time processing with good results. The dipolar representation is designed to deal with the problem of estimating the directionality of data. Experimental results are provided by using the standard PETS dataset and compared with the mixture of Gaussians and background subtraction methods. © 2008 Springer-Verlag.},
     bibsource = {dblp computer science bibliography, https://dblp.org},
     biburl = {https://dblp.org/rec/conf/iberamia/LuqueLDP08.bib},
     document_type = {Conference Paper},
     doi = {10.1007/978-3-540-88309-8_11},
     journal = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
     keywords = {Background model; Background subtraction method; Competitive neural network; Data sets; Hardware implementations; Mixture of Gaussians; Realtime processing; Video segmentation, Hardware; Image segmentation; Pixels, Neural networks},
     source = {Scopus},
     sponsors = {Lisbon University Institute (ISCTE); Fundacao para a Ciencia e Tecnologia (FCT); Asociacion Espanola de Inteligencia Artificial (AEPIA); Associacao Portuguesa para a Inteligencia Artificial (APPIA)},
     url = {https://doi.org/10.1007/978-3-540-88309-8_11},
}

Bibliometric data

The following data has been extracted from resources such as OpenAlex, Dimensions, PlumX or Altmetric.

5

CITATIONS

5 Total citations

0 Recent citations

0.99 Field Citation Ratio

n/a Relative Citation Ratio

A Dipolar Competitive Neural Network for Video Segmentation

Cites

The following graph plots the number of cites received by this work from its publication, on a yearly basis.

Papers citing this work

The following is a non-exhaustive list of papers that cite this work:

[1] T. Bouwmans. “Traditional and recent approaches in background modeling for foreground detection: An overview”. In: Computer Science Review 11–12 (May. 2014), p. 31–66. ISSN: 1574-0137. DOI: 10.1016/j.cosrev.2014.04.001. URL: http://dx.doi.org/10.1016/j.cosrev.2014.04.001.

[2] N. K. Rai, S. Chourasia, and A. Sethi. “An Efficient Neural Network Based Background Subtraction Method”. In: Proceedings of Seventh International Conference on Bio-Inspired Computing: Theories and Applications (BIC-TA 2012). Springer India, Dec. 2012, p. 453–460. ISBN: 9788132210382. DOI: 10.1007/978-81-322-1038-2_38. URL: http://dx.doi.org/10.1007/978-81-322-1038-2_38.

[3] N. K. Rai, A. Singh, and S. A. Mazhari. “Video Segmentation Using Neural Network and Distributed Genetic Algorithm”. In: Proceedings of the International Conference on Soft Computing for Problem Solving (SocProS 2011) December 20-22, 2011. Springer India, 2012, p. 229–237. ISBN: 9788132204916. DOI: 10.1007/978-81-322-0491-6_22. URL: http://dx.doi.org/10.1007/978-81-322-0491-6_22.

[4] K. Ryan, A. Amer, and L. Gagnon. “Spatiotemporal Region Enhancement and Merging for Unsupervized Object Segmentation”. In: EURASIP Journal on Image and Video Processing 2009 (2009), p. 1–13. ISSN: 1687-5281. DOI: 10.1155/2009/797052. URL: http://dx.doi.org/10.1155/2009/797052.