Smart Advertising in Videos Based on Comprehensive Content Analytics
Yi Zhang, Fan Luan, Yu–Gang Jiang · 2019
Traditional video advertising ignores the contents of the video and its relation to the advertisement. In this demo, we develop a smart advertising system based on comprehensive content analytics, which can automatically match the contents of the video and the advertisement to promote the best advertising spaces. Our platform structurizes videos into meta-data with multiple semantic labels and temporal-spatial information. High level scene semantics is extracted based on the meta-data to match the theme of the advertisement. The system has served for industries, analyzed large scale video data and generated millions of advertising opportunities.