The Community for Technology Leaders
2018 IEEE Fourth International Conference on Multimedia Big Data (BigMM) (2018)
Xi'an
Sept. 13, 2018 to Sept. 16, 2018
ISBN: 978-1-5386-5322-7
pp: 1-5
Yumeng Liang , Beijing University of Posts and Telecommunications, Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia, Beijing, 100876, P. R. China
Wu Liu , Beijing University of Posts and Telecommunications, Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia, Beijing, 100876, P. R. China
Kun Liu , Beijing University of Posts and Telecommunications, Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia, Beijing, 100876, P. R. China
Huadong Ma , Beijing University of Posts and Telecommunications, Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia, Beijing, 100876, P. R. China
ABSTRACT
With the rapid growth of online videos, video advertising, as a main source of income for video streaming websites, has attracted increasing attention. It is challenging to insert advertisements at suitable positions while preserving comfortable watching experience of users. Existing methods usually insert advertisements at the fixed positions and neglect the variations of scenes, which can extremely reduce the attractiveness of videos due to intrusion to important visual elements. In this paper, we propose a method to automatically generate and embed appealing textual advertisements for online videos. First we estimate the visual significance of the main elements in the video frames via human face localization and saliency detection. Next we design an efficient algorithm to recognize the scene changes with the visual significance map, through which the system can find stable areas in distinct scenes for advertising. At last, a series of aesthetic designing principles are adopted to generate attractive advertisements which are in harmony with the style of video scenes. User studies show that our system can achieve the best user experience compared with the state-of-the-art methods as well as comparable results with commercial advertisements designed by professional designers.
INDEX TERMS
advertising data processing, face recognition, feature extraction, learning (artificial intelligence), video signal processing, video streaming
CITATION

Y. Liang, W. Liu, K. Liu and H. Ma, "Automatic Generation of Textual Advertisement for Video Advertising," 2018 IEEE Fourth International Conference on Multimedia Big Data (BigMM)(BIGMM), Xi'an, 2018, pp. 1-5.
doi:10.1109/BigMM.2018.8499465
91 ms
(Ver 3.3 (11022016))