Gabor spectrogram based risk of aliasing
Dimitri Van De Ville, Koen N. Denecker, Wilfried R. Philips, Ignace A. Lemahieu · Ghent University Academic Bibliography (Ghent University) · 2000
Resampling is a part of many image processing tasks. Traditional linear resampling techniques must compromise between edge-preservation and anti-aliasing. This paper presents a method to locally estimate the risk of aliasing based on a space-frequency representation by means of the Gabor spectrogram. The proposed approach can be applied to general lattice arrangements of the source and the target lattice. To show the feasibility of this new method, results of a practical application are shown, i.e., resampling an image for gravure printing reproduction by a non-linear edge-preserving resampling technique based on the risk of aliasing. Key-Words: Gabor spectrogram, joint time-frequency analysis, aliasing, non-linear resampling