ON THE EFFICIENCY OF A NEW EFFICIENT ALGORITHM TO COMPUTE THE 2-DIMENSIONAL DISCRETE FOURIER-TRANSFORM

The efficiency claims in a paper by Gertner, 1 based on a Radon transform for reducing the number of 1-D discrete Fourier transforms (DFT's) to compute a 2-D DFT, are offset by a costly increase in the number of additions required to perform the Radon transform. A common factor 2-D fast Fourier transform (FFT) algorithm to compute the 2-D DFT is O(N2 log2 N) computationally, whereas this new approach reduces the number of FFT additions and multiplications by 25%, but adds O(N3) additions. This correspondence narrows the benefit of the original paper to applications where very specialized architectures are used.