Optimal Design Of Laplacian Pyramid Based On The Structural Similarity Index

Laplacian pyramid is a kind of multi-scale image analysis,which has good multiresolution characteristics and has important research significance in the field of image processing.In this paper,several traditional reconstruction algorithms of Laplacian pyramid(LP)are introduced.On this basis,the structural similarity measurement index(SSIM)is used as the optimization criterion to optimize the reconstruction of LP,so as to improve the quality of reconstructed image and enhance the anti-noise performance of LP system.In order to obtain more scale information of signals,LP needs to be decomposed into multiple levels.Therefore,for single-level LP and multi-level LP reconstruction optimization algorithms,the main research work of this paper is as follows:1.Based on the framing pyramid,the optimization design of LP system is carried out with SSIM as the optimization standard.Firstly,the single-stage LP is analyzed,which is equivalent to a set of oversampled filter banks.Under the constraints of LP structure,SSIM is used as the optimization criterion to optimize the filter bank synthesis terminal.The experimental results show that compared with the traditional LP reconstruction algorithm,the reconstruction algorithm based on SSIM not only has better denoising effect,but also has higher SSIM value and better visual perception effect.2.The single-level LP reconstruction optimization algorithm is extended to multilevel LP,and the multi-level LP is analyzed.Firstly,the multistage LP reconstructed end is equivalent to a uniformly decimated tree-shaped filter,and then it is transformed into a uniformly decimated filter bank,and the SSIM-optimized integrated end filter is determined.The simulation results show that the multi-level LP reconstruction algorithm based on SSIM is also better than the traditional reconstruction algorithm in denoising,and when the noise increases gradually,the denoising effect is better than the traditional method.