User Targeted Offline Advertising using Recognition Based Demographics and Queue Scheduling

Abstract

Offline advertisements are static in nature. Advertising companies use billboards for advertising. These billboards display advertisements in a random fashion depending on the investment made by the advertiser. Advertisers pay a fixed amount of money for displaying their advertisements and not on the basis of relevant viewership. The technology proposed in the paper ensures that this disparity is handled wherein offline advertisements are targeted to the relevant audience. The technology has been named TARP which is an abbreviation for Target. Advertise. Revolutionise. Promote. TARP uses built in cameras on offline advertising platforms such as billboards & TV Screens in malls, restaurants, metro & airports to target advertisements based on gender, age and other relevant demographics. The technology is a boon for the advertising industry and benefits both advertisers and viewers. It displays what viewers want to see and who the advertisers want to reach out to. Convolutional neural networks are used to generate demographics of viewing population. Centroids of the viewing population are maintained for each billboard. Advertisements search for the most relevant billboard for display. Display of advertisements is monitored by a queue scheduling algorithm. The research paper proposes an algorithm to generate demographics, search most relevant billboard for each advertisement as well as generate priority queues.