On the calculation of the Minkowski dimension for grayscale images

In the article the result of the statistical experiment is presented. The purpose of the experiment is inspection of the correctness of the algorithm for determining the value and accuracy of Minkowski dimension for grayscale images. For arbitrary image from internet the sample population is constructed. Then for every element of this sample population the Minkowski dimension is calculated by the box counting method. Then for this sample population the mean and standard deviation are defined. Although the box counting method singly has a slow accuracy but in combine with this statistical approximation the Minkowski dimension with the better precision is get. In this article for inspection of the correctness two sample populations are constructed. The first sample population is generated by shifts the image down and the second sample population is generated by shifts image left. Two different sample populations have the little differ means and standard deviation but the Wilcoxon test rejects the hypothesis about the including these sample populations to different general populations. Such result allows make algorithm for more précised the set's Minkowski dimension.