Privacy has received considerable attention but is still largely ignored in the multimedia community. Consider a cloud computing scenario where server resource-abundant, and capable of finishing designated tasks. It envisioned that secure media applications with privacy preservation will be treated seriously. In view fact scale-invariant feature transform (SIFT) been widely adopted various fields, this paper first to target importance privacy-preserving SIFT (PPSIFT) address problem...

The discrete fractional Fourier transform (DFRFT) is the generalization of transform. Many types DFRFT have been derived and are useful for signal processing applications. We introduce a new type DFRFT, which unitary, reversible, flexible; in addition, closed-form analytic expression can be obtained. It works performance similar to continuous (FRFT) efficiently calculated by FFT. Since FRFT generalized into affine (AFT) (the so-called canonical transform), we also extend (DAFT). derive two...

The fractional Fourier transform (FRFT) is a useful tool for signal processing. It the generalization of transform. Many operations, such as convolution, correlation, and Hilbert transform, are defined from it. In fact, FRFT can be further generalized into linear canonical (LCT), we also use LCT to define several operations. this paper, discuss relations between operations described above some important time-frequency distributions (TFDs), Wigner distribution function (WDF), ambiguity (AF),...

The continuous fractional Fourier transform (FRFT) performs a spectrum rotation of signal in the time-frequency plane, and it becomes an important tool for time-varying analysis. A discrete has been developed by Santhanam McClellan (see ibid., vol.42, p.994-98, 1996) but its results do not match those corresponding transforms. We propose new (DFRFT). DFRFT DFT Hermite eigenvectors retains eigenvalue-eigenfunction relation as continous FRFT. To obtain eigenvectors, two orthogonal projection...

The concepts of quaternion Fourier transform (QFT), convolution (QCV), and correlation, which are based on algebra, have been found to be useful for color image processing. However, the necessary computational algorithms their complexity still need some attention. We develop efficient QFT, QCV, correlation. conventional complex two-dimensional (2-D) (FT) is used implement these operations very efficiently. With algorithms, we only two 2-D FTs a six one-side QCV or correlation 12 two-side...

The fractional Fourier transform is a useful mathematical operation that generalizes the well-known continuous transform. Several discrete transforms (DFRFT's) have been developed, but their results do not match those of case. We propose new DFRFT. This improved DFRFT provides similar to and also retains rotation properties.

Removing nighttime haze technique is an important and necessary procedure to avoid ill-condition visibility of human eyes. To increase the scene, there are some existing methods for dehazing in daytime. Specifically, these have a satisfactory quality daytime but most them not all successful nighttime. In this paper, proposed method can be properly applied images even they ill-properties low overall contrast, brightness, “blue shift” airlight color. The basic concept use color transfer [7],...

Objective image quality assessment (IQA) plays an important role in the development of multimedia applications. Prediction IQA metric should be consistent with human perception. The release newest database (TID2013) challenges most widely used metrics (e.g., peak-to-noise-ratio and structure similarity index). We propose a new methodology to build model using regression approach. score is set nonlinear combination features extracted from several difference Gaussian (DOG) frequency bands,...

An adaptive time-frequency representation (TFR) with higher energy concentration usually requires complexity. Recently, a low-complexity short-time Fourier transform (ASTFT) based on the chirp rate has been proposed. To enhance performance, this method is substantially modified in paper: i) because wavelet used for instantaneous frequency (IF) estimation not signal-dependent, ASTFT novel measure addressed; ii) order to increase robustness IF error, principal component analysis (PCA) replaces...

The linear canonical transform (the LCT) is a useful tool for optical system analysis and signal processing. It parameterized by 2/spl times/2 matrix {a, b, c, d}. Many operations, such as the Fourier (FT), fractional (FRFT), Fresnel transform, scaling operations are all special cases of LCT. We discuss eigenfunctions FT, FRFT, have been known, we derive LCT based on these operations. find, different cases, that also forms. When |a+d|<2, scaling, chirp multiplication Hermite functions, but...

This paper is concerned with the definitions of discrete fractional cosine transform (DFRCT) and sine (DFRST). The DFRCT DFRST are based on eigen decomposition DCT DST kernels. same idea as that Fourier (DFRFT); eigenvalue eigenvector relationships between DFRCT, DFRST, DFRFT can be established. computations for even or odd signals planted into half-size calculations. will reduce computational load by about one half.

We present a new approach to the least-squares design of stable infinite impulse response (IIR) digital filters. The is accomplished by using an iterative scheme in which denominator polynomial obtained from preceding iteration treated as part weighting function, and each carried out solving standard quadratic programming problem that yields rational function. When converges, truly solution obtained. method then extended address IIR two-dimensional (2-D) Examples are included illustrate...

An adaptive FIR filter based on the least mean p-power error (MPE) criterion is investigated. First, some useful properties of MPE function are studied. Three main results as follows: 1) a convex coefficients; so it has no local minima. 2) When input process and desired both Gaussian processes, then same optimum solution conventional Wiener for any p. 3) non-Gaussian may have better than solution. Next, (LMP) algorithm derived application examples presented. Consequently, when signal...

This paper presents a new moment-preserving thresholding technique, called the binary quaternion-moment-preserving (BQMP) thresholding, for color image data. Based on representing data by quaternions, statistical parameters of can be expressed through definition quaternion moments. Analytical formulas BQMP thus determined using algebra quaternions. The computation time is order size. By quaternion-moment-based operators are designed application processing, such as compression, multiclass...

A technique for suppressing the transient states of IIR notch filter is investigated. This uses vector projection to find better initial values filters. When a or comb used eliminate power line (AC) interference in recording electrocardiograms (ECG), performance with suppression than that conventional arbitrary condition. The improvements this are at cost additional computation load beginning filtering.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML"...

Many useful relations between the Gabor transform (GT) and fractional Fourier (FRFT) have been derived. First, we find that, like Wigner distribution function (WDF), FRFT is also equivalent to rotation operation of GT. Then, show that performing scaled inverse (IFT) along an oblique line GT f(t) can yield its FRFT. Since closely related FRFT, use it for analyzing characteristics Compared with WDF, does not cross-term problem. This advantage important applications filter design, sampling,...

Space shift keying (SSK) modulation and its extension, the generalized SSK (GSSK), present an attractive framework for emerging large-scale MIMO systems in reducing hardware costs. In SSK, maximum likelihood (ML) detector incurs considerable computational complexities. We propose a compressed sensing based detector, NCS, by formulating SSK-type detection criterion as convex optimization problem. The proposed NCS requires only O(n <sub xmlns:mml="http://www.w3.org/1998/Math/MathML"...

Conventional Fourier analysis has many schemes for different types of signals. They are transform (FT), series (FS), discrete-time (DTFT), and discrete (DFT). The goal this article is to develop two absent fractional methods. proposed methods (FRFS) (DTFRFT), they the generalizations (FS) respectively.

Digital signal and image processing using reduced biquaternions (RBs) are introduced in this paper. RBs an extension of the complex numbers, following doubling procedure. Two useful representations (e/sub 1/-e/sub 2/ form matrix representation) discussed Besides, we propose a new representation (the polar form) to calculate multiplication conjugation easily. Furthermore, define unique suitable RB norm its conjugate. These definitions similar compatible with numbers. The efficient algorithms...

This paper is concerned with the definitions of discrete fractional Hartley transform (DFRHT) and Fourier (DFRFT). First, eigenvalues eigenvectors matrices are investigated. Then, results eigendecompositions used to define DFRHT DFRFT. Also, an important relationship between DFRFT described, numerical examples illustrated demonstrate that proposed a better approximation continuous than conventional defined Finally, filtering technique in domain applied remove chirp interference.

In this paper, a new comb filter design method using fractional sample delay is presented. First, the specification of transformed into that design. Then, conventional finite impulse response (FIR) and allpass techniques are directly applied to with specification. Next, we develop constrained approach improve performance direct method. Finally, several examples an experiment power line interference removal in electrocardiogram (ECG) signal demonstrated illustrate effectiveness approach.

A new technique for IIR multiple notch filter design is proposed. The specification of first transformed into that allpass filter. Then, we develop an effective approach to this desired realization proposed equivalent the Due mirror-image symmetry relation between numerator and denominator polynomials filter, can be realized by a computationally efficient lattice structure with very low sensitivity. Moreover, some examples are presented examine effectiveness method.

Based on discrete Hermite–Gaussian-like functions, a fractional Fourier transform (DFRFT), which provides sample approximations of the continuous transform, was defined and investigated recently. In this paper, we propose new nearly tridiagonal matrix, commutes with (DFT) matrix. The eigenvectors matrix are shown to be DFT eigenvectors, more similar Hermite–Gaussian functions than those developed before. Rigorous discussions relations between eigendecomposition newly proposed described....