Tech

This AI finds you posing for Instagram right where you took the photo

Anyone would say that we live in a world like Minority report, but with examples like this, we are more and more convinced. Software and artificial intelligence continue to take leaps and bounds, and the imagination of many developers usually reaches ridiculous heights that give life to projects as interesting as they are terrifying like the one we are telling you about today. Do you feel watched? We present to you The Follower.

Do not pose, they are watching you

Instagram AI public cameras

You love it pose for an instagram photo. Admit it, absolutely nothing happens. But if you thought that no one saw that placement of legs, hips out and pouts in the air, it is because you did not know the existence of The Follower. Created by Dries Depoorter, this search software is responsible for finding the exact moment an Instagram photo was taken in a specific place in the world.

This sounds like complete science fiction, and although the description may be too generous, the truth is that it complies with what is announced in certain situations.

Twitter user image

Dries DePoorter

@driesdepoorter

🔥NEW PROJECT!🔥
‘The Follower’ is software searching how an Instagram photo was taken with the help of AI and open cameras.

Project page: https://t.co/Djzlu3rf39
YouTube video: https://t.co/WpD3gdkPIu
Support my work: https://t.co/lP6wVXSkJ3

🧵👇 https://t.co/MEQ9kwz8mY

September 15, 2022 • 18:16

There are eyes all over the world

Security cameras

The secret is to be able to see what was happening in the place of the photo. To do this, the software The Follower uses public webcams that broadcast live what is happening on the streets. So for example we can see what is happening in Times Square, or in a busy corner of London. The software works with this valuable source of resources, which is dedicated to comparing the video files with Instagram captures made in said areas.

The scrutiny of the photos is quite simple to do using the geolocated tags or the manual tags in the description of the photo, but the complicated thing comes later, and that is that you have to review weeks of recording and find exactly the person who was passing by and that matches the appearance of the original Instagram photo.

The-Follower-Dries-Depoorter-07.gifThe-Follower-Dries-Depoorter-05.gif

Where’s Wally in easy mode

As you can see, the results are surprising, since the software’s AI is capable of recognizing very precise details of the clothing of the person who appeared in the photo. Thanks to the result, it is quite fun to see the behind-the-scenes process, since we can see how on many occasions the photo requires several attempts, postures and help from other people to make the final shot.

The current limitations of the system is that the photo is required not to have been published or taken a long time ago, since it only saves a few weeks of video of the recordings taken by public cameras. In other words, to find you on your vacation a year ago, it would take a whole year to store recordings, not to mention the analysis time that the AI ​​would require to find you among so many resources.

Is this type of software dangerous?

Instagram location

Surely the first thing you will have thought is that this type of software violates the privacy of users and threatens their security. Well, not exactly. The photos that have been used have been uploaded by users who have a public Instagram account and who have geotagged themselves in that place to indicate where they took the photo, so they are the ones who are sharing their privacy. Another matter is that of the public cameras, which are there recording images non-stop without the consent of those who walk by. Both sources of information are defenseless separately, but combined they can offer very specific data.

For example, we could know who took the photo or if the person was accompanied by other people at the time. The possibilities are endless, and the fear is even greater. What’s your opinion about it?

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *