FIELD: computing technology.
SUBSTANCE: method for selection and demonstration of contextual information associated with a television video stream includes receiving data of the video stream containing the video intended for translation on the screen of the user, analyzing the data, performing a search request for contextual information, receiving a search result in form of images, text, hyperlinks, advertising materials, selecting search results in accordance with the data of the user profile, demonstrating the selected information on the screen of the user, wherein the analysis of the video stream data begins with video-analytical extraction of graphic objects from the video stream, executed using neural networks, wherein the features of each object are identified, the time of presence of the object on the screen is monitored, a search request for contextual information for the object is performed while said object is present on the screen for more than 5 seconds, the received search results are analyzed and divided by data type into two groups, the search results for each object are demonstrated alternately at a predetermined step from the first and second groups.
EFFECT: provided is effective demonstration of a larger amount of contextual information of advertising nature while maintaining the attention of the user.
1 cl, 1 dwg
