Fractional-order orthogonal Chebyshev Moments and Moment Invariants for image representation and pattern recognition. (February 2019)

• Record Type: Journal Article
• Authors: Benouini, Rachid; Batioua, Imad; Zenkouar, Khalid; Zahi, Azeddine; Najah, Said; Qjidaa, Hassan
• Is Part Of: Pattern recognition. Volume 86(2019:Feb.)
• Page Start: 332
• Page End: 343
• Publication Date: 2019-02
• Subjects: Fractional-order orthogonal moments -- Fractional-order Chebyshev polynomials -- Moment invariants -- Image representation -- Pattern recognition -- Fast and accurate computation; FCP Fractional-order Chebyshev Polynomials -- FCM Fractional-order Chebyshev Moments -- LM Legendre Moments -- GM Geometric Moments -- GegM Gegenbauer Moments -- FGM Fractional-order Geometric Moments -- FCMI Fractional-order Chebyshev Moment Invariants -- FGMI Fractional-order Geometric Moment Invariants -- LMI Legendre Moment Invariants -- GegMI Gegenbauer Moment Invariants -- DTMI Discrete Tchebichef Moment Invariants -- DKMI Discrete Krawtchouk Moment Invariants -- GMI Geometric Moment Invariants -- MSE Mean Squared Error -- RST Rotation, Scaling and Translation -- ROI Region Of Interest; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2018.10.001
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: Introducing a new set of fractional-order moments for image representation. Constructing a set of fractional-order moment invariants for pattern recognition. Discussing the computational aspects of the proposed fractional-order moments and moment invariants. Providing numerical experiments to demonstrate their validity and superiority. Abstract: In this paper, we present a new set of fractional-order orthogonal moments, named Fractional-order Chebyshev Moments (FCM). We initially introduce the necessary relations and properties to define the FCM in the Cartesian coordinates. Then, we provide the theoretical framework to construct the Fractional-order Chebyshev Moment Invariants (FCMI), which are invariants with respect to rotation, scaling and translation transforms. In addition, we devoted a substantial attention to enhance their computational time and numerical accuracy. Consequently, the numerical experiments are carried out to demonstrate the validity of the introduced fractional-order moments and moment invariants in comparison with the classical methods, with regard to image representation capability and object recognition accuracy on several publicly available databases. The presented theoretical and experimental results demonstrate the efficiency and the superiority of the proposed method.