anonymous video analytics (AVA)
Anonymous video analytics (AVA) is a strategy for gathering metrics about digital signage audience engagement.
In the past, marketers could only gather data about how many people passed by a digital sign. The problem with this method is that many people pass by a digital sign without looking at it.
Anonymous video analytics uses camera sensors and a program for analyzing mathematical algorithms to detect a passing face that turns towards the sign. Unlike facial recognition software, AVA does not match a particular face to a face in a database. (This is what makes it anonymous.) Instead, AVA relies on a technique known as pattern detection. The AVA program consults a database of pixel patterns to determine when something in front of the sign is a human face looking in the right direction. In very simple terms, the program looks for patterns such as dark pixels where eyes should be and lighter pixels where cheeks should be.
Once the AVA program has determined that the pixel patterns are of a human face, the program can record how long the face looked at the sign. It may also categorize the face according to demographic groups, such as gender and age.
Example video from Intel Cognovision Solution