Construction of Optimal Multiwavelet And Its Application in Image Denoising

摘要

In the application of multiwavelet in image processing, the properties such as orthogonality, symmetry, vanishing moment, regularity, optimum time-frequency resolution (OPTFR) are very important. Through the algorithm we designed that synthetically used the method of parameterization, genetic algorithm and compound Simpson formula, we constructed multiwavelets which are better than the present ones in those properties. Application in image processing showed that the denoising performance of some of our multiwavelets are better than the best multiwavelets of Jiang's J_EX2 and CL when dealing with smooth image and texture image.