The Efficiency of Images Reduction Algorithms with Small-Sized and Linear Details
Kirill Smelyakov, Anton Datsenko, Vitalii Skrypka, Aganur Akhundov · 2019
The article describes the most popular algorithms for reducing the scale of a digital image, experimenting with applying these algorithms to an image with linear (objects without entrails) and small-sized objects for the most used scale reduction coefficients, analyzing the effectiveness of the experimental results and developing practical recommendations for applying the zoom algorithms to an image with a large number of small parts. In this experiment were used the following algorithms: brightness of the center pixel, the average value, brightness closest to the average value, truncated mean, the median number of brightness, the arithmetic average on the set of some count of brightness. Results of these algorithms were compared by the coefficient of homogeneity and the coefficient of contrast parameters and were presented in tables. At the end of the experiment, it was clarified that the best type of convolution for all types of images and scale reduction coefficients doesn't exist. Different algorithms are used for different types of images.