Multi-strategy-based novel advertisement discovery method, electronic device and readable storage medium

Abstract

The present application provides a new advertisement discovery method based on multiple strategies, the method comprising: extracting advertisement segments from the sample database; extracting the voiceprint features of each advertisement segment according to the broadcast time of each advertisement segment; Use the advertising voice template to match the voiceprint features of each advertising segment one by one to select unknown advertising segments from the advertising segments. The advertising voice template is generated based on the known voiceprint features; use the voice recognition model to calculate each unknown advertising segment The sound boundary time point; the unknown advertisement segment is segmented according to the sound boundary time point to obtain the unknown segment; the strategy feature attribute information of each unknown segment is calculated respectively. The beneficial effects of the present invention mainly lie in that: the application uses a machine computing model to intelligently calculate the strategy characteristic attribute information of the advertisement segment, and accurately calculates the new advertisement in the advertisement segment of the radio and television program, thus shortening the advertisement sample production cycle and avoiding artificial intervene.

Description

technical field [0001] The invention relates to the technical field of advertisement supervision, in particular to a multi-strategy-based new advertisement discovery method, electronic equipment and a readable storage medium. Background technique [0002] With the development and opening up of my country's radio and television industry, TV stations have more and more TV programs. As the main source of income for TV stations, TV advertisements have put forward higher requirements for TV station program management. Advertising supervision is the most important link. Moreover, with the continuous deepening of the field of radio and television content management, the rapid expansion of various digital audio and video content in the advertising samples of radio and television production at all levels, advertising management occupies a large proportion and workload in the television industry. Requirements for chemical retrieval. [0003] Since the supervision of advertisements in r...

AI Technical Summary

Problems solved by technology

[0003] Since the supervision of advertisements in radio and television programs mainly relies on making advertisement samples and comparing the characteristics of radio and television programs, so as to achieve the purpose of monitoring and supervision, at present, it mainly relies on manual marking of video recordings, and monitors and monitors the characteristic attributes of advertisement videos and voices to start monitoring and supervision of advertisements. The time point and the end time point are intercepted and saved. This method needs to browse a large number of videos. Radio and TV advertisements have the characteristics of many programs, strong periodicity, and strong repetition. Therefore, manual screening has a heavy workload and repetitive work, which is very inconvenient. question

Images