| ABSTRACT: | THE FAST ALGORITHM STUDY FOR WAVELET IMAGE DENOISING BASED ON DSP
H. Zhicheng, Z. Jiyin, G. Shuang, L. Yu
Wavelet transform is a effective way to digital signal processing. Author propose a fast algorithm for image
denoising on the basis of discrete wavelet transform. Projecting image signal to the two-dimentional discrete
wavelet field, and using the half-soft threshold method to deal with wavelet coefficient and achieve the goal
of denoising. Compared with traditional wavelet algorithm, this one, based on Mallat tower algorithm, adopt
a fast algorithm of wavelet decomposition and reconstruction as well as improved FFT. It reduce operation
amount by a large margin, and raise the algorithmic parallel degree. The experimental result of DSP
emulation indicates this algorithm has better real-time character, has offered a new method for digital image
processing.
|