{"id":24724,"date":"2018-06-26T10:25:12","date_gmt":"2018-06-26T14:25:12","guid":{"rendered":"https:\/\/www.tun.com\/blog\/?p=24724"},"modified":"2022-03-16T10:49:36","modified_gmt":"2022-03-16T14:49:36","slug":"mit-artificial-intelligence-see-people-through-walls","status":"publish","type":"post","link":"https:\/\/www.tun.com\/blog\/mit-artificial-intelligence-see-people-through-walls\/","title":{"rendered":"Artificial Intelligence Can See People Through Walls"},"content":{"rendered":"

In a groundbreaking new project, MIT researchers have developed a computerized system that uses artificial intelligence (AI) to <\/span>see people through walls<\/span><\/a>. <\/span><\/p>\n

\u201cRF-Pose,\u201d as they have dubbed the technology, functions as real-life X-ray vision. <\/span><\/p>\n

The technology uses a neural network to analyze radio frequencies that reverberate off people\u2019s bodies. This allows the system to detect people\u2019s postures and movement in real time, even from behind walls or in the dark. RF-Pose then creates a two-dimensional stick figure that moves as the person does.<\/span><\/p>\n

\"\"<\/p>\n

So, what inspired the team to develop the technology? <\/span><\/p>\n

\u201cEstimating the human pose is an important task in computer vision with applications in surveillance, activity recognition, gaming, etc.,\u201d said <\/span>Dina Katabi<\/span><\/a>, the Andrew & Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT\u2019s <\/span>Computer Science and Artificial Intelligence Laboratory (CSAIL)<\/span><\/a>, who led the research.<\/span><\/p>\n

The research was presented on June 21 at the <\/span>Conference on Computer Vision and Pattern Recognition (CVPR)<\/span><\/a> in Salt Lake City, Utah. <\/span><\/p>\n

How does the technology work?<\/b><\/h2>\n

AI learns by example. That is, by showing a neural network a large data set of items, it will learn to identify certain trends in the data set. <\/span><\/p>\n

For RF-Pose, the researchers had to teach AI to associate particular radio signals with specific human actions.<\/span><\/p>\n

To do so, they collected thousands of images of people doing activities like walking, talking, standing, sitting, opening doors, opening elevators, and more. <\/span><\/p>\n

They used these images to create stick figures, posing just as the people in the images did. <\/span><\/p>\n

They paired these stick figure poses with corresponding radio signals and showed them to the AI. <\/span><\/p>\n

\u201cThis combination of examples enabled the system to learn the association between the radio signal and the stick figures of the people in the scene,\u201d Katabi said. <\/span><\/p>\n

\u201cPost-training, RF-Pose was able to estimate a person\u2019s posture and movements without cameras, using only the wireless reflections that bounce off people\u2019s bodies.\u201d<\/span><\/p>\n

RF-Pose was never explicitly trained to identify people\u2019s actions through walls. But, because it is trained to identify people\u2019s movement based on radio signals, barriers don\u2019t affect the way it can detect movement.<\/span><\/p>\n

In the video below, RF-Pose was shown monitoring the movement of people through walls. It is capable of tracking the movements of multiple people at one time. <\/span><\/p>\n