Image Compression by Vector Quantization with Recurrent Discrete Networks

 

Abstract

In this work we propose a recurrent multivalued network, generalizing Hopfield’s model, which can be interpreted as a vector quantifier. We explain the model and establish a relation between vector quantization and sum-of-squares clustering. To test the efficiency of this model as vector quantifier, we apply this new technique to image compression. Two well-known images are used as benchmark, allowing us to compare our model to standard competitive learning. In our simulations, our new technique clearly outperforms the classical algorithm for vector quantization, achieving not only a better distortion rate, but even reducing drastically the computational time. © Springer-Verlag Berlin Heidelberg 2006.

Cites

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

Citation

Please, cite this work as:

[Lóp+06] D. López-Rodríguez, E. M. Casermeiro, J. M. Ortiz-de-Lazcano-Lobato, et al. “Image Compression by Vector Quantization with Recurrent Discrete Networks”. In: Artificial Neural Networks - ICANN 2006, 16th International Conference, Athens, Greece, September 10-14, 2006. Proceedings, Part II. Ed. by S. D. Kollias, A. Stafylopatis, W. Duch and E. Oja. Vol. 4132. Lecture Notes in Computer Science. cited By 6; Conference of 16th International Conference on Artificial Neural Networks, ICANN 2006 ; Conference Date: 10 September 2006 Through 14 September 2006; Conference Code:68317. Athens: Springer, 2006, pp. 595-605. DOI: 10.1007/11840930_62. URL: https://doi.org/10.1007/11840930_62.

BibTeX

@InProceedings{LopezRodriguez2006a,
     author = {Domingo López-Rodríguez and Enrique Mérida Casermeiro and Juan Miguel Ortiz-de-Lazcano-Lobato and Ezequiel López-Rubio},
     booktitle = {Artificial Neural Networks - {ICANN} 2006, 16th International Conference, Athens, Greece, September 10-14, 2006. Proceedings, Part {II}},
     title = {Image Compression by Vector Quantization with Recurrent Discrete Networks},
     year = {2006},
     address = {Athens},
     editor = {Stefanos D. Kollias and Andreas Stafylopatis and Wlodzislaw Duch and Erkki Oja},
     note = {cited By 6; Conference of 16th International Conference on Artificial Neural Networks, ICANN 2006 ; Conference Date: 10 September 2006 Through 14 September 2006; Conference Code:68317},
     pages = {595–605},
     publisher = {Springer},
     series = {Lecture Notes in Computer Science},
     volume = {4132},
     abstract = {In this work we propose a recurrent multivalued network, generalizing Hopfield’s model, which can be interpreted as a vector quantifier. We explain the model and establish a relation between vector quantization and sum-of-squares clustering. To test the efficiency of this model as vector quantifier, we apply this new technique to image compression. Two well-known images are used as benchmark, allowing us to compare our model to standard competitive learning. In our simulations, our new technique clearly outperforms the classical algorithm for vector quantization, achieving not only a better distortion rate, but even reducing drastically the computational time. © Springer-Verlag Berlin Heidelberg 2006.},
     bibsource = {dblp computer science bibliography, https://dblp.org},
     biburl = {https://dblp.org/rec/conf/icann/Lopez-RodriguezCOL06.bib},
     document_type = {Conference Paper},
     doi = {10.1007/11840930_62},
     journal = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
     keywords = {Algorithms; Learning systems; Mathematical models; Recurrent neural networks; Vector quantization, Classical algorithm; Hopfield’s model; Sum-of-squares clustering; Vector quantifier, Image compression},
     source = {Scopus},
     url = {https://doi.org/10.1007/11840930_62},
}