Study Of Image Denoising Using Multi-wavelet Transform

In this paper,the method of image threshold denoising is investigated by using the 2-band multi-wavelet with multiplicity two.Above all, we discuss the development of multi-wavelet theory ,the course of study and the superiority of the image denoising method are analysing by using the multi-wavelet trasform.Next,it has introduced the basics of the multi-wavelet, the algorithm features of the multi-wavelet transform are analysed,and the concrete realization algorithm of multi-wavelet image denoising is introduced in detail.One after another,the emphases of this paper is threshold denoising to image,which will use the way of milti-wavelet transform. In this paper,the denoising modle is the additive white gaussian noise image. According to the characteristics of multi-wavelet transform and feature of studied image ,it analyzes the threshold denoising method of image by the use of multi-wavelet.When we apply the multi-wavelet transform to image threshold denoising,it analyse the way of choosing threshold and threshold function,and discusses its importance.After a series of study of hard and soft threshold functions,we found that they are all have flaws.Based on the hard and soft threshold functions, we put forward a new threshold function to get over the flaws in this paper,it will less the bias and variance between original image and denoised image.Eventually, we choose the 256 gray-scale image as the object of the study to conduct experiment in MATLAB. Image denosing is achieved by using the methods of neighborhood averaging,median filter,wavelet threshold function and multi-wavelet threshold function.According to the above study, we compare the quality standards and visual effects of the original image with the denoised,it mainly explains the follwing points.First,the superiority of wavelet denoising method.Second,we select CL3 multi-wavelet as a multi-wavelet bases, multi-wavelet method is superior than the wavelet method.Third,we still select CL3 multi-wavelet as a multi-wavelet bases,the feasibility and superiority of the new threshold denoising.