An Intelligent Pixel Replication Technique by Binary Decomposition for Digital Image Zooming

Image zooming is the process of enlarging the spatial resolution of a given digital image. We present a novel technique that intelligently modifies the classical pixel replication method for zooming. Our method decomposes a given image into layer of binary images, interpolates them by magnifying the binary patterns preserving their geometric shape and finally aggregates them all to obtain the zoomed image. Although the quality of our zoomed images is much higher than that of nearest neighbor and bilinear interpolation and comparable with bicubic interpolation, the running time of our technique is extremely fast like nearest neighbor interpolation and much faster than bilinear and bicubic interpolation.