Kurt R.A. Giambastiani

The eye is an incredible organ, but it has its limits. We can’t see infrared or ultraviolet, much less anywhere else along the massive electromagnetic spectrum. We can’t see very well at night or when we’re surrounded by a lot of reflected sunshine. The eye is particularly susceptible to defects, from myopia and astigmatism to cataracts and floaters–not unexpected in what is essentially a high-precision organic instrument, but definitely a limitation. Aeons ago, for example, someone with vision like mine would have had the nickname, “Food for Wolves.”

Another limitation of our otherwise remarkable eyes is that we can’t see subtle, minute shifts in light or color. Our eye (or, more accurately, our brain) averages them out, giving us a more stable view of the world than one in which we see the quaking of each individual leaf in a gentle breeze or the shudder of everything in the room as a heavy truck rolls by outside.

But a team of scientists at MIT have found that where our eyes fail us, a computer can give us a helping hand.

The process, known as Eulerian Video Magnification, is a computer program that analyzes these subtle shifts in color and/or motion and amplifies them. Viewing a video that has been modified in this way, you can see–in real-time–the pulse of a newborn or whether or not a sleeping child is actually breathing.

There’s a blog post and a quick video showing examples of the results, and while it’s just a start, it opens up a whole range of possibilities for diagnostic and analytical applications. Already it can be used to check the heart rate of extremely premature babies, for whom physical contact brings risks. Other ideas spring to mind, such as monitoring movement in large structures or imaging changes in air density and temperature. This technique can literally show us what is invisible to us, but is right before our eyes.

What impressed me the most about this team, though, was the fact that they’ve put all their code out there, online, for anyone who wants to use it (in a non-commercial application). You see, this doesn’t require million-dollar labs or cutting edge equipment. A video camera and a computer is pretty much all you need (a nerdy aptitude would help, for sure).

Let’s hear it for the geeks!

k