A Fast Image Enhancement Method by Virtual Image Pyramid Sequence Fusion
Xi Dai · Chinese Journal of Computers · 2014
To accelerate the implementation of the image enhancement method,we proposed a new global-to-local strategy for fast image enhancement.Based on this practical strategy,a scalable and fast image enhancement framework is proposed to construct an implementation of the existing image enhancing filter in O(N)time using the virtual image pyramid sequence fusion.Under this framework,the input image is first remapped globally by a series of tentative remapping functions to generate a virtual candidate image sequence,which is then recombined locally to a single output image under a special edge-aware pixel-level fusion rule.To avoid halo artifacts,both the output image and the virtual candidate image sequence are transformed into multi-resolution pyramid representations.Application on the local Laplacian filter indicates that the proposed framework can induce a very fast approximation of this edge-preserving filter.This fast approximate method is able to generate very close results to the original local Laplacian filter and achieves about two orders of magnitude speedup.