crazy process called Eulerian Video Magnification
So how does it work? It picks up on the very slight nuances in a video that you can’t detect, such as the way a face reddens as blood is pumped through the body. It grabs these visualisations from a video sequence and applies spatial decomposition and temporal filtering to the frames. Then it amplifies the colour so these nuances become amazingly dynamic and easy to see.
Of course, this X-ray vision-style technique has a lot of possible medical applications — it can automatically pick up vital signs from standard videos, track the way a baby is breathing, or analyse the way an artery is pulsing. It could perhaps also be used for very close surveillance, which is kind of a scary thought. We might see some of this played out in the near future, because the MIT team said it’s planning to release the code soon. [MIT via BuzzFeed]
GIF images: John Herrman/BuzzFeed