Adaptation Neighborhoods of Self-Organizing Maps for Image Restoration

Adaptation neighborhoods of self-organizing maps for image restoration are presented in this study. Generally, self-organizing maps have been studied for the ordering process and the convergence phase of weight vectors. As a new approach of self-organizing maps, some methods of adaptation neighborhoods for image restoration are proposed. The present algorithm creates a map containing one unit for each pixel. Utilizing pixel values as input, the image inference is carried out by self-organizing maps. Then, an updating function with a threshold according to the difference between the input value and the inferred pixel is introduced, so as not to respond to a noisy input sensitively. Therefore, the inference of original image proceeds appropriately since any pixel is influenced by surrounding pixels corresponding to the neighboring setting. Experimental results are presented in order to show that the present methods are effective. Key-Words: Self-organizing maps, Image restoration, Topological neighborhood, Learning Peak signal to noise ratio, Numerical experiment