A Hybrid Denoising Model Basing On Total Variation and Analysis in Transform Domain
Qiuqi Ruan · Signal Processing · 2008
A novel hybrid denoising model which combines Total Variation and wavelet shrinkage basing on function space is pro- posed in this paper.The hybrid model denoises by minimizing Total Variation under the regularization constraint in small function space; it preserves edge feature and weakens staircasing.There is no Gibbs phenomenon in the denoised image and the novel model can restore lost information.DCT is used to characterize texture;it is introduced into the hybrid model in this paper,and get another hybrid denoising model which preserves texture better.Both theoretical analysis and experiments have verified the validity of the new model.