Can AI deliver tailored billboard advertising?
Wouldn’t it be cool if a computer could assess your parent’s mood and signalize when’s the best time to approach them if you want to use their car?
Thanks to a new piece of analysis software from the Fraunhofer Institute, this might soon be possible.
Object recognition using AI is a relatively simple task nowadays. For example, OpenCV, a software library that software developers use for computer vision applications, can be combined with deep learning AI frameworks to quickly build applications that can detect faces and people. But determining the mood displayed on the face of a person is a little more challenging. Thankfully the SHORE® face analysis software developed by Fraunhofer makes this possible, thus opening a multitude of exciting new possibilities. By detecting facial features, such as the eyes and mouth, and determining whether they are open or closed, the AI estimates whether the subject is happy, sad, surprised or angry. Additionally, SHORE® is even able to determine the gender of the person with an accuracy of over 90%, and even estimate their age to within 7 years. So, where could this be used?
Traditional advertising has lost a lot of ground to online advertising and, despite electronic displays at bus stops and underground stations, it can’t compete with the precision and relevance offered by online advertisers. If electronic advertising displays could assess the age and gender of their brief audience, it might be possible to deliver more relevant content to those looking. Data privacy laws would seem in the way of such a development, but the Fraunhofer team have managed to squeeze this technology into a camera platform, named AVARD, making it redundant to offer up data to the cloud for analysis, as the entire AI runs inside the camera. AVARD has also attained a “Privacy Seal” from ePrivacy, thus confirming its data privacy credentials.
The facial expression recognition capability of SHORE® could also be used in vehicles to determine the mood of the driver, perhaps recommending a break when detecting fatigue, or sending out a caution if signs of discomfort are visible. Perhaps, with time, our computers will develop a little emotional intelligence.