Interpolation of 2-D signal by subsequence FFT

Discrete interpolation between successive samples of a sequence is often required in digital signal processing. The subsequence approach for 1-D interpolation is extended to two dimensions to avoid redundant operations. An improved intermediate sequence is proposed to preserve the Hermitian symmetry in interpolating real-valued signals. The resulting algorithms is very efficient and convenient because it permits the use of inverse FFTs with sizes that are the same as the corresponding forward transforms. >