17-03-2017, 05:01 PM
Image elimination methods are often based on minimizing an appropriately defined energy function. Many gradient-dependent energy functions, such as the Potts model and the total variation of degradation, regard the image as a constant function in parts. In these methods, some important information such as edge sharpening and location is well preserved, but some detailed image feature as texture is often compromised in the process of downsizing. For this reason, we propose a method of image elimination based on local adaptive regularization, which can adaptively adjust the degree of degradation of the noisy image by adding the term of fidelity to the spatial variable, in order to better preserve the characteristics of Fine scale of the image. Experimental results show that the proposed recess elimination method can achieve a leading-edge subjective visual effect, and the signal-to-noise ratio (SNR) is also objectively improved by 0.3-0.6 dB.
Edge preserving Denoising
Denoising is a critical step in many image processing tasks. Linear methods have been very popular for their simplicity and speed, but their use is limited because they tend to blur the images. Nonlinear methods require more time, but in general they are much better.
Our group conducts research on different methods of nonlinear elimination. The key idea is to perform the anisotropic diffusion as opposed to the isotropic diffusion performed by linear methods. Nonlinear methods behave differently depending on the content of the image. Near the edges diffusing along the edges, but not through and in the smooth areas perform the standard isotropic diffusion. Thus, nonlinear methods eliminate noise and simultaneously preserve edges.