Digital Signage Personalization Through Analysis of the Visual Information About Viewers

Digital signage has been gaining more and more popularity. However, personalization of digital signage is still challenging and it is getting even more difficult with the appearance of low regulating the ways of personal data storage and processing. Most advertisement frameworks can detect general information about viewers such as age and gender that is not enough to recognize his/her real interests. There are efforts aimed at detection of emotions and prediction interests based on the mood as well as usage of modern 3D cameras for collecting distance to the viewers and their heights. Also, the topic of detection logotypes in images is getting more popular. In this paper, we are researching possibilities of integrating possible information collection techniques for better prediction of viewers' interests. These are integrated into the concept of the system that contains human detection, analytics, advertisement, and statistic modules.