Image re-sampling from a perspective of wave field propagation

This paper deals with the problem of 2D image resampling. Image re-sampling refers to the process of re-sizing an image at a higher or lower resolution one. This process has often been tackled in 2D spatial domain and seems to still remain there up to now. In this paper, we extend image resampling to a higher domain by considering it as a wave field propagation problem in 3D space. By doing so, we observe that the possibility is more increased for a re-scaled image to preserve spectral information well in it compared to previous interpolation techniques which tend to usually sacrifice high frequency information when the image is down-scaled or upscaled also. We present a novel 3D space based image resampling method by using a technique of wave field propagation. Experiments results on both natural and CG images are provided which validates our method. Comparison results with a few representative interpolation techniques clearly show desirable characteristics of our 3D wave field based image re-sampling over those in 2D domain.