Image comparison algorithm based on the analysis of pairs of positions and values of vector elements

The task of image comparison is very important in the development of intelligent systems based on image recognition. The article considers the possibility of using a new algorithm developed by us, which determines the correlation (similarity) of two vectors. The main idea of the comparison was that two vectors are similar if the difference between the positions and values of the elements in the original and sorted variants is small enough. The article also presents the results of comparing images, in particular face images, based on this algorithm. However, this method can be used for other comparable objects represented as vectors. The original version of this article supplied to AIP Publishing contained an error in step 9 of the algorithm used to compare vectors A and B. This error has been corrected in the new article published on 19 September 2024.