An evolutionary approach for gray-level image zooming

Image resizing is widely applied in many fields such as medical image processing, consumer electronics and space application. In this paper, a two-phase adaptive image zooming method for gray-scale image magnifying is presented. For each local area under processing, the first work is trying to find a best-matched remote window within the image based on the structural similarity. The second step is then applying the relationship of pixels in the remote window for evolving interpolation functions, which is then used to calculate the expected values of pixels for filling the enlarged grid. The interpolation function varies as the sample window moving across the whole image. Experiment results show that the proposed method is superior to pixel replication and bilinear techniques in terms of visual quality.