Adaptive experiences based on face analysis are as old as humanity, and they don’t require face recognition

There is an increasing public concern around face recognition and the inherent violation of privacy it represents. In a world where cameras are omnipresent, as we walk through cities, we have literally nowhere to hide. With face recognition, we are left in a condition of complete passivity, we cannot opt-out. Once our face is part of a database, then it’s potentially there forever… we can change a hacked password, but we cannot change our face. Such considerations, as well as the more worrying implementations of face recognition that are being introduced in countries such as China that are eerily reminiscent of the dystopian classics, add to the widespread concern with the technology.

Since the technology developed by MorphCast includes a camera and is based on face analysis, people who first hear about our product sometimes associate it with face recognition. This impression, however, is based on a fundamental misunderstanding of what our product does. So let’s set the record straight.

Face analysis does not imply face recognition

MorphCast tracks emotions and other basic features through face analysis to adapt contents: it does not identify individuals, and does not store any sort of facial fingerprint. According to MorphCast CEO Stefano Bragagni, MorphCast “does the same kind of thing we do when we are on the streets, on a train, or in a bar: we look at strangers, and we evaluate them based on their appearance, their apparent mood, their clothing, their posture, and then we instinctively adapt our behaviour to the people around us. On first contact we can now use machine learning to adapt to their expressions, mood and reactions.”

Facial analysis is so embedded in human interactions that we are not even fully aware we do it. We naturally respond differently, based on instant facial analysis of our interlocutor: is it an adult or a child? a man or a woman? are they smiling or visibly stressed? We, as humans, instinctively adapt our behaviour depending on such signals.

MorphCast’s concept of adaptive media brings this important component of human interaction to our screens.

Similarly to what happens in our everyday experience, adapting responses to who’s in front of a screen does not need to come at the cost of privacy. If you took a picture of every single person you meet on the streets, take notice of where they go, and then go up to them and say: “oh, I know you were here before and I know what you like to eat for lunch”, well, that would be creepy. This is in many ways what is happening today on the internet, with advertisers using various tracking methods to follow our every move.

On the contrary, MorphCast’s approach resembles more closely what happens in real life: we look at people, take an adequate response based on who’s in front of us, and then move on. We adapt constantly to who’s in front of us, but we don’t need to identify who’s in front of us, nor to take notes about each person we cross on the street. In real life we can have perfectly adequate and reasonably customised responses without the need to break anonymity or invade other people’s privacy.

MorphCast does exactly the same: thanks to live facial analysis, it does not need to identify users or to employ invasive tracking techniques. On the contrary, it requires your deliberate, conscious consent to enable the webcam. Your image is then processed directly inside the browser, and who’s in front of the webcam at any given point in time is what informs your adaptive media experience.

These are some of the features that differentiate clearly MorphCast’s adaptive and privacy-respective experience from facial recognition techniques or legacy approaches to micro-targeting and tracking.

To see how MorphCast can track your mood in real-time without sending your personal data to a server try: https://morphcast.com/static/spectrum/

To see how MorphCast technology can be used to create rich adaptive media experiences, try our full MorphCast demo at: https://www.morphcast.com