Fuzzy colour naming based on sigmoid membership functions

Robert Benavente*, Maria Vanrell

*Corresponding author for this work

    Research output: Chapter in BookChapterResearchpeer-review

    6 Citations (Scopus)

    Abstract

    In this paper, we present some improvements towards a computational model for colour naming. Our model is based on fuzzy set theory and each colour category is considered a fuzzy set with a characteristic function. In previous works, we had proposed a model based on the use of a Sigmoid-Gaussian as membership function for the chromatic categories. Although it provided good results, the Sigmoid-Gaussian model presents some drawbacks due to the parameter dependence between the Sigmoid and the Gaussian functions. To overcome this, we propose two new functions which are based only on products of Sigmoids avoiding the problems introduced by the Gaussian function. The results obtained by the new functions are compared to the previous ones. Although the improvement in terms of the fitting error is not very significant, the new functions show a higher degree of adaptability which will allow improving the modelling of the whole colour naming space. The functions are also used to label the Munsell colour array and the new membership functions provide similar categorizations than real observers.

    Original languageEnglish
    Title of host publicationProceedings of the 2nd European Conference on Colour in Graphics, Imaging, and Vision (CGIV'2004)
    Place of PublicationSpringfield, Va. (US)
    PublisherSociety for Imaging Science and Technology
    Pages135-139
    Number of pages5
    Edition1
    ISBN (Print)089208250X, 9780892082506
    Publication statusPublished - 1 Jan 2004

    Publication series

    NameCGIV 2004 - Second European Conference on Color in Graphics, Imaging, and Vision and Sixth International Symposium on Multispectral Color Science

    Fingerprint

    Dive into the research topics of 'Fuzzy colour naming based on sigmoid membership functions'. Together they form a unique fingerprint.

    Cite this