SmartScreen: Slimme digitale schermen voor doelgerichte advertenties en communicatie
In an attempt to make online advertisements more relevant and hence more effective, advertisers increasingly personalize advertisements based on the viewer’s physical characteristics and/or online behavior. In this project, we investigate whether the concept of online personalized advertisements can also be applied in offline settings. In particular, we explore whether digital billboards can be ‘enhanced’ with person and object detection algorithms to adapt the message on the screen to the person in front of it (e.g. if a male person stands in front of the screen, it shows an advertisement for male’s clothes). The question remains whether adapted advertisements are more effective: Is there an increase in advertising impact and is it sufficiently large for companies to invest in this technology? Furthermore, it is important to confirm that the technology is in accordance with the current privacy legislation (AVG or GDPR), and whether people accept the technology as part of their daily lives.
There are three main parts in this project: (1) algorithm development, (2) measuring advertisement impact of adapted messages, and (3) legislation and public perception. Concerning the development of the algorithms (1), we collaborate with engineers from EAVISE (KU Leuven). Making use of recent technologies such as deep learning, these engineers design software that reliably detects physical characteristics. In part (2), we investigate whether consumers consider advertisements adapted to their physical characteristics more relevant than advertisements that are not adapted, and whether improved relevance actually translates into increased attention to and recall of the message. In order to provide an accurate image of visual attention to the advertisement, we make use of mobile eye-tracking devices. Finally (3), it is explored whether the current technology of ‘adapted advertisement based on the input of a camera and an algorithm’ is in accordance with privacy legislation such as AVG (GDPR). In addition, we also study how people perceive this technology with regards to privacy, since accordance with the law does not always translate into accordance with the public.
- Connect (BE)
- Haystack nv (BE)
- KU Leuven (BE)
- Museum M (BE)
- Profacts (BE)