Gaussian Pyramid

MultiresolutionImageProcessing1)GaussianPyramid: Image pyramids, as depicted in Figure 2, are a fundamental concept in multi-rate image processing. A well-known example is the Gaussian pyramid, which is made up of low-pass filtered, downsampled images of the preceding stage of the pyramid, where the base stage g0 is defined as the original image g0(x) = f(x). Higher stages are defined by gs(x) = (ξgs−1(x)w(x,ξ), where w(x,ξ) is a weighting function that is identical for all stages, termed the generating kernel. Most practical approaches, however, stop before reaching the top of the pyramid. After processing the images at each stage, the output is reassembled by fusing together images of successive stages in a reconstruction step.