Frit Denoising Method Based On Wavelet Transform Matrix

This paper had discussed the FRIT denoising methods using wavelet transform matrix.Using wavelet transformation matrix to process images aim at Mallat algorithm, the method feasible,has good restructuring. Furthermore, constructed wavelet vector,and used the wavelet transform matrix for the FRIT image denoising.FRIT for image denoising exists three aspects:First, FRIT is based on the finite Radon transform,and FRAT image processing require for the prime number size, for the average size of the square of the binary image, the image must first be extended to the size of prime numbers, and then to FRIT;Secend, would be introduced in the image expanding random noise; Three,FRAT modulo the definition of a straight line in the image into the "surround" effect, to a certain extent, offset the advantages of FRIT.Meanwhile, both the reconstruction and decomposition of the wavelet transform need to progressively calculated,that if each of the calculation through a matrix-vector multiplication implementation, will be very convenient, and when the filter length is greater than the signal length, the signal needs to be extension.In this paper, Firstly for the traditional FRIT can only process a limited number of prime-size images, to seek an improved methods based on wavelet transform matrix FRIT to adapt to a binary pixel image; Secondly, in order to solve the introduction of random noise in image processing problems, presents a progressive image/column block processing method; Thirdly, by way of the progressive introduction of the three kinds of boundaries based on Mallat algorithm for extension methods, and use the three kinds of continuation method applied to the proposed improved image denoising; Finally, proposed a ridge of a wavelet transform method to solve the problem of "around".These methods of improved based on wavelet transform matrix, interlocking with each other, the former is the basis of the latter, the latter is the former kind of supplement, gradually improved. Experiments show that a clear image for linear feature has a good de-noising effect, greatly reducing the boundary effect.