A critical review of the existing methods of object selection in a video stream

The selection of contours of objects in the video stream in intelligent video surveillance systems is one of the main operations in image processing for further analysis, since the contour contains all the necessary information for recognizing objects by their shape. The purpose of this article is to give a critical overview of existing methods for selecting objects in a video stream. The article is also devoted to the analysis of image recognition methods and their search in the video stream. The evolution of the structure of convolutional neural networks used in the field of diagnostics of computer video streams is analyzed. An overview of existing methods for selecting objects in the video stream is made. It is worth saying that the recognition of the entire object only by its contour allows you not to consider the internal points of the image and thus significantly reduce the amount of information processed, providing the opportunity to analyze images in real time. In this article, based on available sources and published literature, answers are given to questions about what is the methodology and methodology of scientific research. So, the article considers the problem of selecting objects in the video stream in the tasks of detecting alarming events by intelligent video surveillance systems. As a result of the work, the author came to the conclusion that a promising direction for further research is the development of algorithms for selecting the contours of images of objects that implement a two-scale statistical image model. The practical significance of the article lies in the fact that, in order to improve the basic characteristics of intelligent video surveillance systems, algorithms are proposed for selecting the contours of images of objects necessary to ensure the identification of four types of alarming events: the appearance and location of an object in the observation zone, the movement of an object in a prohibited direction, the abandonment of an object and the overturning of an object.