No abstract

No abstract

The Fourier transform of a continuous function, evaluated at frequencies expressed in polar coordinates, is an important conceptual tool for understanding physical continuum phenomena. An analogous tool, suitable for computations on discrete grids, could be very useful; however, no exact analogue exists in the discrete case. In this paper we present the notion of pseudopolar grid (pp grid) and the pseudopolar Fourier transform (ppFT), which evaluates the discrete Fourier transform at points of the pp grid. The pp grid is a type of concentric-squares grid in which the radial density of squares is twice as high as usual. The pp grid consists of equally spaced samples along rays, where different rays are equally spaced in slope rather than angle. We develop a fast algorithm for the ppFT, with the same complexity order as the Cartesian fast Fourier transform; the algorithm is stable, invertible, requires only one-dimensional operations, and uses no approximate interpolations. We prove that the ppFT is invertible and develop two algorithms for its inversion: iterative and direct, both with complexity O(n 2 log n), where n × n is the size of the reconstructed image. The iterative algorithm applies conjugate gradients to the Gram operator of the ppFT. Since the transform is ill-conditioned, we introduce a preconditioner, which significantly accelerates the convergence. The direct inversion algorithm utilizes the special frequency domain structure of the transform in two steps. First, it resamples the pp grid to a Cartesian frequency grid and then recovers the image from the Cartesian frequency grid.

One of the major challenges related to image registration is the estimation of large motions without prior knowledge. This paper presents a Fourier-based approach that estimates large translations, scalings, and rotations. The algorithm uses the pseudopolar (PP) Fourier transform to achieve substantial improved approximations of the polar and log-polar Fourier transforms of an image. Thus, rotations and scalings are reduced to translations which are estimated using phase correlation. By utilizing the PP grid, we increase the performance (accuracy, speed, and robustness) of the registration algorithms. Scales up to 4 and arbitrary rotation angles can be robustly recovered, compared to a maximum scaling of 2 recovered by state-of-the-art algorithms. The algorithm only utilizes one-dimensional fast Fourier transform computations whose overall complexity is significantly lower than prior works. Experimental results demonstrate the applicability of the proposed algorithms.

We introduce a multiscale scheme for sampling scattered data and extending functions dened on the sampled data points, which overcomes some limitations of the Nyström interpolation method. The multiscale extension (MSE) method is based on mutual distances between data points. It uses a coarse-to-ne hierarchy of the multiscale decomposition of a Gaussian kernel. It generates a sequence of subsamples, which we refer to as adaptive grids, and a sequence of approximations to a given empirical function on the data, as well as their extensions to any newlyarrived data point. The subsampling is done by a special decomposition of the associated Gaussian kernel matrix in each scale in the hierarchical procedure.

No abstract

Abstract. In this paper we propose a direct method for the solution of the Poisson equation in rectangular regions. It has an arbitrary order accuracy and low CPU requirements which makes it practical for treating large-scale problems.The method is based on a pseudospectral Fourier approximation and a polynomial subtraction technique. Fast convergence of the Fourier series is achieved by removing the discontinuities at the corner points using polynomial subtraction functions. These functions have the same discontinuities at the corner points as the sought solution. In addition to this, they satisfy the Laplace equation so that the subtraction procedure does not generate nonperiodic, nonhomogeneous terms.The solution of a boundary value problem is obtained in a series form in O(N log N ) floating point operations, where N 2 is the number of grid nodes. Evaluating the solution at all N 2 interior points requires O(N 2 log N ) operations.Key words. boundary value problem, Poisson equation, rectangular region, spectral method, corner discontinuities, polynomial subtraction AMS subject classifications. 35J05, 45L10, 65P05PII. S1064827595288589 1.Introduction. An important step in the development of fast numerical solvers for elliptic equations in complicated domains is an algorithm for the solution of boundary value problems for constant coefficient elliptic equations in rectangular regions. After solving a nonhomogeneous equation with some "convenient" boundary conditions, one must solve, in a correction step, the homogeneous problem with specified (Dirichlet or Neumann) boundary conditions. As a resulting boundary distribution may not satisfy the required compatibility conditions at the corner points, singularities may arise in the process of the solution, thus destroying the convergence rate even if the final solution is smooth.When the computational region is discretized on a N × N grid using some loworder (finite difference or finite element) scheme, the resulting system of linear algebraic equations is represented by a sparse matrix. Such a matrix can be inverted in O(N 2 ) (or O(N 2 log 2 N ) arithmetic operations using a "fast solver" [4]. However, since the method is of low order, the resolution N must be very large if a high accuracy is desired.Application of high-order (pseudo) spectral methods, based on global expansions into orthogonal polynomials, e.g., Chebyshev polynomials, to the solution of elliptic equations, results in a full matrix problem. The cost of inverting such a matrix using the best current algorithms is O(N 3 ) operations [2]. Besides, the accuracy decreases considerably as the dimension N 2 of the system grows due to accumulation of roundoff errors.When using the Chebyshev method, the fast computation of the expansion coefficients requires that the problem be discretized on a nonuniform grid. For timedependent problems, when elliptic equations arise due to a time-discretization pro-

The goal of still color image segmentation is to divide the image into homogeneous regions.Object extraction, object recognition and object-based compression are typical applications that use still segmentation as a low-level image processing. In this paper we present a new method for color image segmentation. The proposed algorithm divides the image into homogeneous regions by local thresholds. The number of thresholds and their values are adaptively derived by an automatic process, where local information is taken into consideration. First, the watershed algorithm is applied. Its results are used as an initialization for the next step, which is iterative merging process. During the iterative process regions are merged and local thresholds are derived. The thresholds are determined one-by-one at different times during the merging process. Every threshold is calculated by local information on any region and its surroundings. Any statistical information on the input images is not given. The algorithm is found to be reliable and robust to different kind of images.