Image Denoising Using Weight Model

The article deals with the issues of primary image processing in computer vision-based systems. The main approaches to solving the problems of image enhancement and restoration, including using wavelet transformations, are described. Filtering algorithms based on threshold processing of the detailing coefficients at various levels of the wavelet decomposition are considered. An approach is proposed based on the construction and analysis of a weight model that reflects the importance of image elements in terms of their contribution to the total energy. A method for constructing a multiscale weight model, a procedure for removing noise in an image using it, and the results of experiments on processing noisy test images are given. It is shown that using the proposed procedure, it is possible to ensure the removal of Gaussian noise with a large standard deviation value.