Thesis (Ph. D.)--Oregon State University, 1974. / Typescript (photocopy). Includes bibliographical references. Also available on the World Wide Web.

No description available.

This dissertation considers a method for processing two-dimensional (2-D) signals (e.g. imagery) by transformation to a coordinate space where the 2-D operation separates into orthogonal 1-D operations. After processing, the 2-D output is reconstructed by a second coordinate transformation. This approach is based on the Radon transform, which maps a two-dimensional Cartesian representation of a signal into a series of one-dimensional signals by line-integral projection. The mathematical principles of this transformation are well-known as the basis for medical computed tomography. This approach can process signals more rapidly than conventional digital processing and more flexibly and precisely than optical techniques. A new formulation of the Radon transform is introduced that employs a new transformation--the central-slice transform--to symmetrize the operations between the Cartesian and Radon representations of the signal and to aid in analyzing operations that may be susceptible to solution in this manner. It is well-known that 2-D Fourier transforms and convolutions can be performed by 1-D operations after Radon transformation, as proven by the central-slice and filter theorems. Demonstrations of these operations via Radon transforms are described. An optical system has been constructed to derive the line-integral projections of 2-D transmissive or reflective input data. Fourier transforms of the projections are derived by a surface-acoustic-wave chirp Fourier transformer, and filtering is performed in a surface-acoustic-wave convolver. Reconstruction of the processed 2-D signal is performed optically. The system can process 2-D imagery at approximately 5 frames/second, though rates to 30 frames/second are achievable if a faster image rotator is added. Other signal processing operations in Radon space are demonstrated, including Labeyrie stellar speckle interferometry, the Hartley transform, and the joint coordinate-frequency representations such as the Wigner distribution function. Other operations worthy of further study include derivation of the 2-D cepstrum, and several spectrum estimation algorithms.

No description available.

No description available.

No description available.

In this thesis, we present a new class of algorithms that determine fast Fourier transforms for a given finite group G. These algorithms use eigenspace projections determined by a chain of subgroups of G, and rely on a path algebraic approach to the representation theory of finite groups developed by Ram (26). Applying this framework to the symmetric group, Sn, yields a class of fast Fourier transforms that we conjecture to run in O(n2n!) time. We also discuss several future directions for this research.

A discrete Fourier transform, or DFT, is an isomorphism from a group algebra to a direct sum of matrix algebras. An algorithm that efficiently applies a DFT is called a fast Fourier transform, or FFT. The concept of a DFT will be introduced and examined from both a general and algebraic perspective. We will then present and analyze a specific FFT for the symmetric group.

The ability to efficiently and sparsely represent seismic data is becoming an increasingly important problem in geophysics. Over the last thirty years many transforms such as wavelets, curvelets, contourlets, surfacelets, shearlets, and many other types of ‘x-lets’ have been developed. Such transform were leveraged to resolve this issue of sparse representations. In this work we compare the properties of four of these commonly used transforms, namely the shift-invariant wavelets, complex wavelets, curvelets and surfacelets. We also explore the performance of these transforms for the problem of recovering seismic wavefields from incomplete measurements.

No description available.