DEVELOPMENT AND RESEARCH OF THE MODEL FOR VIDEO INFORMATION CLASSIFICATION

The article is devoted to solving the scientific problem of classifying video content in faceof an increase information volume. Computer vision a very relevant field artificial intelligencetechnologies application expand capabilities various search and archive systems. Theauthors give definitions main terms studied subject area. A formalized statement theproblem be solved presented. detailed classification possible options for problemis given. With rapid development technology, digital showing anexplosive growth trend. sports videos great importance archiving digitalcontent on server. Many data mining machine learning algorithms have made strides inmany areas (such as classification, regression, clustering). However, most thesealgorithms common drawback when training test samples are same featurespace follow distribution. This discusses problemof automatic annotation, also develops modelbased deep big data. As part this study, authors developed model that improvesthe quality which improves results. results computationalexperiment show proposed can effectively used classify eventswithin area based use convolutional neural network. At sametime, high accuracy provided. Compared with other models,the has advantages simple implementation, fast processing speed, highclassification accuracy, generalization ability.