A new spatially variant apodization (SVA) algorithm that uses a 3/4 filled aperture prior to two dimensional discrete Fourier transform (2-D DFT) to form the image. The algorithm can be used, for example, to improve contrast and resolution on synthetic aperture radar (SAR) imagery, with a lower degree of oversampling (and thus, fewer pixels) than other algorithms require. This can translate into more efficient use of radar displays and processor memory. Additional efficiencies of memory and computing power may be realized when Automatic Target Recognition (ATR) algorithms operate on this imagery. Embodiments of this invention use convolution kernels at two different spacings, which are better tuned to the local phase relationships of mainlobe and sidelobes with a 3/4 filled aperture. As such, these embodiments suppress sidelobes without sacrificing resolution, at an aperture-filling ratio of 3/4, rather than 1/2, as is usually used.