Adaptive Denoising Algorithm Based on Nonlocal Means in Image Processing

The article is devoted to the existing problem associated with the absence of resource-lights denoising algorithms, capable of producing good-quality output images in the different intensity noise conditions without blurring the boundaries, contours, and basic structure. The adaptive algorithm suggested in the article allows solving this problem due to the developed algorithms of splitting the search region into two sets of similar and points different from the pixel and adapting of the kernel type to the image region, depending on the presence or detection of structural and smooth pixels. The results of the suggested algorithm and the standard method of nonlocal means are compared with the metrics of the peak signal-to-noise ratio and structural similarity. It is found out that the developed adaptive algorithm surpasses the standard method both on numerical results and on the quality of the image processing.