Scientists at MIT are using AI to develop X-Ray vision-like technology

• The new system can track human movement behind walls.
  
• Currently, the system can estimate movements with 83% accuracy.
  
• Radio waves are being used to detect movements and postures.

Scientists from the Massachusetts Institute of Technology (MIT) in the US, have developed an artificial intelligence (AI) based X-ray vision system that can track and sense movements of people and their postures. And what’s exciting about this technology is that it can track human movements through walls.

So, our dream to have X-Ray vision may soon be a reality.

How does it work?

The new project, which is being called the “RF-Pose”, uses AI to track people using a wireless device. The model works by creating stick figures of people's movement in real time, by analysing radio signals that bounce off people’s bodies.

Researchers are using thousands of gathered images of people during different activities, like sitting, walking or opening a door to train the AI model. Each of the gathered photos is then shown to a camera that creates a stick figure. This is then analysed by the model’s neural network along with the corresponding radio signals.


Where can it be used?

Scientists believe that the technology has multiple applications in medical sciences, including monitoring people suffering from Parkinson's disease and Multiple Sclerosis(MS). The new tracking system will be able to provide a better understanding of how the disease progresses and will also allow doctors to adjust medications according to the data analysed.

This technology can even be used by the military and special forces to conduct rescue missions and help locate lost survivors.

It can even be integrated into multiplayer gaming platforms, which will enable users to move around in the house or a gaming arena to create new experiences. For example, laser tag could be played in a virtual environment with this tech tracking each player providing much realistic gaming experience.

Future of the Tech

Currently, the tech has an accuracy of 83% to identify movements. After the devices training is complete, the wireless device, RF-Pose, would be able to estimate a person’s posture and movement with complete accuracy by just using wireless reflections of radio waves.

Researchers are working to improve the system. As of now, the model gives just a 2D stick figure representation, but the team is working towards creating 3D figures that would enable them to even identify micromovements.