Finding Video Variations
Place your hands in the right spots, and it is pretty easy to feel your pulse pushing blood through your body. Look at the same spot though, and you probably will not be able to see it. This is because the variations involved are too minute and fast for the human eye and brain to discern, but a new algorithm from MIT can capture all the changing detail.
Originally the researchers were developing the algorithm just to identify color changes within a video, or between multiple still images, but they quickly discovered it could also detect motion. They then developed an algorithm specifically for detecting the variation of motion. This algorithm makes even minute motion perceptible, including the one's pulse in their wrist or the subtle breaths of premature infant. If the motion of a person's pulse cannot be seen, the heart beat can still be seen by monitoring the coloring of the skin as it flushes with fresh blood being pumped to it, and pales as the blood moves away. When compared to a more traditional medical monitor, the pulse detected by the algorithm being run on video from a regular camera (meaning not a high-speed camera) was very accurate.
Perhaps the most important use of this technology will be in medicine, for monitoring systems that do not need to touch the person. Premature babies, for example, are often fragile so even attaching sensors can be risky. Being able to just set up a camera would be much easier and safer.