Image Denoising Algorithm Study Based on Morphological Component Analysis
Cui Da-yan · Computer Knowledge and Technology · 2013
Due to signal acquisition and transmission. Random noise has a significant influence on signal, and even reduces the signal quality. The traditional denoising method can't automatically make the optimal choice between denoising and preserving.This paper demonstrates a relatively lossless way to remove noise from the useful signal. This model designed by this paper is based on such hypothesis: The original image signal is composed of random noise and useful signals, which are morphologically different. Based on this difference, these two components can be sparsely represented in separate dictionary. The next step is to ex tract each component and denoise. Finally, reconstruct the signal with all components while all noise is removed.