NUFFTNon-Uniform Fast Fourier Transform
References in periodicals archive ?
Considering the computational cost [14] and result of energy accumulation [17], we perform NUFFT corresponding to [a.
where [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] denotes NUFFT operator with respect to [(nT).
The computational cost is reduced a lot compared with ML method and MDCFT due to the obtained search step length and NUFFT.
Step 2: Use the inverse NUFFT method to obtain the result of a target at an azimuth angle and a given elevation.
Alternatively, a 1D NUFFT routine of NER type can be used to achieve the values of the filtered projections directly at the non-uniform sampling points of interest [10, 11].
The NUFFT algorithm employed in this paper is that developed in [9], which is based on the use of Kaiser-Bessel interpolation windows, see also [10, 11].
The main idea of the NUFFT algorithm is to approximate the "nonuniform" exponential exp(-j2[pi][y.
In this paper, we substitute the Stolt interpolation and IFFT with the 3D inverse NUFFT to reconstruct the image.
Here, the NUFFT algorithm developed by Fessler and Sutton is implemented.
The overall operation count per NUFFT is O(JM) + O(K[log.
Equation (8) can be computed by the 2D NUFFT algorithm [20] with complexity of O([[mu].
iv) Apply the 2D NUFFT algorithm to compute Equation (8).