Surveillance cameras are everywhere these days: outdoors and indoors, airports, train stations, offices and shops. We can't escape the watchful eye of the camera, even in the wild.
For most surveillance systems, the video is recorded in cycles "just in case". Lately, it has become more common for this video to be sent to various data analysis systems, therefore, it could be used to monitor some specific people.
It goes without saying that Big Brother can violate our privacy. You may not like it, but there is nothing illegal in this story.
What would you say if every detail of your private lifewas collected by different companies and one day came to light on the Internet?
The laws of most countries still do not really suppress the use of facial recognition for commercial purposes, just as it is not forbidden, for example, to photograph people on the street.
But to understand how this situation can be improved, let's look at two approaches to modern image analysis methods.
The first approach is based on comparing certain indicators that are transferable from an image to a ready-made database. The indicators can be: the distance between the eyes, the results of nose measurements, assessments of the shape of the lips and the like.
This approach is similar to identifying a person by their fingerprints. The fingerprints are taken in advance and stored in a database, and when necessary, identification is performed. So, the prerequisite for recognizing a face is good quality photos (full-face and with good lighting) of the right people.
Where can these photos be found? The sources can be different. Maybe you had to look into a camera to create a discount card, or maybe someone could have scanned some documents with your image.
It is relatively easy to deceive the classic facial recognition system. The easiest way is to lower your head and not look at the camera. Most indicators can only be measured from a certain full-face angle. If you wear a baseball cap on your head, then the cameras located higher up (usually installed somewhere quite high) become completely useless.
Some experts advise making a face when passing by a lens. It may work quite well, but it attracts a lot of attention. A pair of sunglasses may be all you need.
The advantage of sunglasses is that they cover the eyes, which are one of the most useful areas on a person's face for recognition systems. Algorithms are not that advanced and cannot handle this situation easily enough. So, large opaque glasses are a challenge for classic systems, as are mirrored models that blind the camera with reflected light.
The second approach to human recognition, which is actively being developed, for example, by Facebook and Google, works in a different way. It will be based on machine learning algorithms and automatic data sampling and will be scaled up to compare with all available electronic sources.
This is something much more flexible that is much harder to deceive. Even with an oxygen mask covering your face, I can’t guarantee that you will remain unknown.
They can use all available data for recognition: the shape of your foot or a hairless spot, your tattoo, your beard, your clothes, and so on. An experimental development by Facebook showed that it can identify a face from any angle with 83% accuracy with a sufficient number of images.
The key point here is a sufficient number of images for comparison. If, to consider the other side, there is only one image of you, even a high-quality one, then the probability of successful recognition drops sharply. This is why Big Data technologies and fast Internet search algorithms have come to the fore.
And here comes the tricky part: should we have openly posted pictures of ourselves online? We can stand by the fact that Facebook or Google uses them for their own marketing purposes, as we can’t hide from these “Big Internet Brothers” anywhere at all. However, nothing prevents a company from digging up the required data online, as long as it is freely accessible.
Let’s say your Facebook is closed with the “Friends Only” setting. What about your random pictures in various other people’s posts? What about your LinkedIn profile? It is very difficult to block all sources, even with complete abstinence from social networks.
The solution to this is still unclear. Probably, there may be a stricter regulation of the biometric data market on the part of the government and a more conscious attitude from society.
So, it's time to get used to the idea that our personal images are pretty much the same thing as a private document or a scan of our credit card. Showing them everywhere is not recommended!
