The underlying principle remains the same as before but the complexity increases with the RF-Capture. As the name suggests, the RF-Capture is installed on the opposite wall and is instructed to send wireless radio frequency signals (Wi-Fi) across a room. When the signals hit any obstacle such as the human body they are reflected and are captured by the RF-Capture. The RF-Capture takes some time before creating the silhouette because as the person walks across the room only certain parts of his/her body reflect the signal. The RF-Capture captures multiple snapshots of the person within a certain period of time and by using a reconstruction algorithm developed by the team members generates a rough figure of the person from which we can see physical features such as head, chest arms and feet. RF-Capture can also show how a person is standing in the room.
The most amazing thing about the RF-Capture is the capability of motion capture through walls. For example if a person faces the wall on the opposite of which the device is installed, he/she can move their hands and this movement is accurately picked up by the device. The team tested the capabilities of the RF-Capture by pitting it against a Kinect that was installed inside the room and found that the RF-Capture was just 2cms off in detecting the movement of the hand which is pretty impressive considering the fact that it is working through a wall.
The team is pretty optimistic about the future of the RF-Capture which could be used for motion sensing games and to monitor movement of people in a room remotely. The RF-Capture can be fine-tuned to detect a person’s breathing and heart rate which would be critical in providing health care services. The team at Gizmodo raised some questions about the privacy issues regarding RF-Capture but the team said that they have included encryption in this from the get go and will be putting measures to prevent malicious use of their invention. For more about the RF-Capture you can visit the project page or watch the short video below.