Studying the effect of paralleling settings on the functioning of a barcode recognition app

The paper proposes a parallel video stream processing method for a linear barcode recognition. The researchers have simulated the system for various memory usage modes. A number of experiments have been carried out to identify the weaknesses and peculiarities of the system. Based on these data, improvements were made to solve the problem of a more efficient identifying of the objects in a video stream. Theoretical studies were carried out on the basis of the parallel computing theory, vector algebra and the pattern recognition theory. The approach enables a more efficient usage of processor capabilities.