Google Mirror Move corresponds to your position on the index of 80,000 photos in real time

This site may generate affiliate commissions from the links on this page. Terms of use.

Google has invested a lot in artificial intelligence to power products like Google Photos and Google Assistant These applications are impressive, but they are not really "fun". Google has just launched a new AI experience entirely geared towards fun. The new Mirror Move experience allows you to identify your pose and match over 80,000 images of other people to show you something. One in a similar position.Why so you can do GIFs, of course.

Although this specific Google experience does not have a practice it's a stunt of the buzzwords of technology. Mirror Move uses artificial intelligence, machine learning, neural networks, augmented reality and much more. You can try it now as long as your computer has a webcam. Just go to the Move Mirror website above and give access to your camera to get started.

Mirror Move works best if you stand far enough away from your computer so that all your joints are in the frame. It is also limited to one person at a time. When you move, the AI ​​badyzes your body and estimates where your joints are. It then matches your poses to a catalog of 80,000 photos of people. Thus, your live camera stream is on one side of the screen, and Mirror Move fills the other with corresponding images in real time. You can even create a GIF file of this process to share it on the Internet.

All this is only for fun, but the technology behind Mirror Move could have many applications. Google calls it PoseNet, and you can learn more about the details in an average post from earlier this year. Like many Google technologies, PoseNet is powered by a convolutional neural network. The camera's flow is routed to the network, which identifies people and maps 17 tracking points to the image. The network combines these points with its catalog of 80,000 images, and you get the output.

PoseNet works in both one-person and multi-person detection modes. There is a separate demo that shows how well it works, but all you get is the 17-point detection wireframe. Mirror Move is an implementation of the single-person version because it is faster and the image library is made up of individual people.

PoseNet could possibly find use in games, fitness tracking, and even interactive art installations. This is not a tool to recognize who people are, so it's less a matter of privacy than tools that can recognize and remember faces.