Computer Vision for Nondestructive Quality Control of Superconducting Materials in Medical Practice

Abstract

In the course of the experimental research, a universal COMSRL program for processing optical source images, including the possibility of streaming data analysis, has been developed. The mathematical model of the program is also developed. Operations of mathematical morphology can be applied to digital image processing in various fields and are particularly useful in medicine. It is important to choose the right shape and size of the structural element, and then you can get great benefit from this method of image processing. The universal COMSRL program has been tested in laboratory experiments. It is recommended to use it in related fields of activity, not only in medicine. The authors present an algorithm for gradient optimization of target functions based on adaptive estimates of lower order moments. The proposed method is simple to implement, computationally efficient, requires little memory, is invariant to diagonal scaling of gradients, and is well suited to problems with large amounts of data and/or parameters.