Двовимірна систолічна структура для обробки матриці елементів дискримінантних функцій

Abstract

The paper examines the features of systolic architecture and analyzes its prospects for parallel processing of matrix elements of linear discriminant functions in pattern recognition tasks. The obtained results indicate the feasibility of using systolic structures for processing matrices of elements in classification tasks and their potential application in artificial neural networks. This architecture can be implemented on programmable logic devices, enhancing the speed of information processing for data classification in intelligent systems.