Search

In this study, we focus on the extension of the trigonometric F-transform for functions in one variable to: (i) a larger domain, (ii) a higher degree of the F m -transform, and (iii) many-variable functions to improve its approximation properties over the entire domain and especially at its boundaries. In addition, the properties of approximation and convergence of direct and inverse extended and multidimensional trigonometric F m -transforms are discussed. Then, direct formulas for partial derivatives of functions of several variables are obtained in terms of trigonometric F m -transforms, which are used to solve the Cauchy problem for the transport equation. A new image compression method is proposed and compared with well-established compression methods such as JPEG, JPEG 2000 and their less complex variations JPEG(APDCBT), JPEG(APUBT3), APUBT3-NUP, and JPEG-FT. We have shown that this t F ¯ 11 -transform image compression method has high accuracy and reasonably low (irreducible) complexity. [ABSTRACT FROM AUTHOR]

Fuzzy transforms (F-transforms) have shown to be a powerful technique to deal with several tasks of image processing and time series. However, F-transforms have an important handicap in order to be applied to data in a broader sense: They are by definition applicable only to functions. This paper presents an extension of the definition of F-transforms to a more general structure of data than functions. Moreover, we provide a theoretical study by showing how the fundamental properties of F-transforms are extended to this new approach. Finally, we include also a section to explore potential applications of this novel extension. [ABSTRACT FROM AUTHOR]