Image Representation Via a Finite Radon Transform
IEEE Transactions on Pattern Analysis and Machine Intelligence
Orthogonal discrete radon transform over pn × pnimages
Signal Processing - Special section: Advances in signal processing-assisted cross-layer designs
The discrete periodic Radon transform
IEEE Transactions on Signal Processing
Generalised finite radon transform for N x N images
Image and Vision Computing
Hi-index | 0.00 |
This paper presents a Discrete Radon Transform (DRT) based on congruent mathematics that applies to N × N arrays where $N \in \mathcal{N}$. This definition incorporates and is a natural extension of the more restricted cases of the finite Radon transform [1] where N must be prime, the discrete periodic Radon transform [2] where N must be a power of 2, and the DRT over pn [3], where N must be a power of a single prime. The DRT exactly and invertibly maps a 2-D image to a set of 1-D projections of length N. Projections are found as the sum of the pixels centred on a parallel set of discrete lines. The image is assumed to be periodic and these lines wrap around the array under modulo N arithmetic. Properties of the continuous Radon transform are preserved in the DRT; a discrete form of the Fourier slice theorem applies, as does the convolution property. A formula is given to find the projection set required to be exactly invertible for arrays with N any composite number, as well as a means to determine the level of redundancy in sampling that is introduced on such composite arrays.