Image denoising study based on balanced multiwavelet

摘要

The particular noise in image not only lower the quality of image but also make denoising more difficult. The multiwavelet with the orthogonality and symmetry decompose the image signal more precisely and denoise better than wavelet. But for multiwavelet transform, the choice of prefilter is extremely important, meanwhile leads to the complication of computation. Furthermore, the image decomposition based on balanced orthogonal multiwavelet is excellent, and does not need pre-filtering. So, in this paper, we propose a method of combining threshold-selecting model with semi-soft thresholding function for image denoising based on the coefficients of balanced orthogonal multiwavelet transform. The simulation results show that this method is superior to others.