Deva Ramanan is “Brilliant”

Phuc Pham | New University

Deva Ramanan, an associate professor here at UC Irvine, was recently honored in Popular Science magazine as one of their “Brilliant 10” young scientists of the year.

The honor was bestowed upon Dr. Ramanan as he became one of the leading pioneers in Computer Vision, a new field of study that focuses on getting computers to “see” the world the way we see it.

Humans see the world in a variety of images and shapes, and we can analyze these constant streams of images to understand what is going on in the world around us. For example, if we see someone falling, our brains tell us that this person may be in danger or may need some type of assistance, and we know how to act from there.

However, computers don’t process these images the same way we do. After watching someone fall, they usually need to be told what to do  and how to react. They can’t think for themselves the way humans can.

Dr. Ramanan is working on a way to fix this. He has contributed heavily to the creation of an algorithm that can process a vast number of situations like falling and use that information to get computers to react properly.

Old methods required the computer to run each situation (millions of them) with a variety of different reactions, wasting valuable time. Now, the process takes only a couple of seconds, with minimal effort on the computer’s part.

Essentially, it works by identifying the more moveable parts of the body (areas like the joints) and then assigning these parts to specific places on an immense array that the algorithm can access when it analyzes an image. Usually, past approaches have tried to analyze larger pieces of the body (the entire arm, or the torso for example) but that was unwieldy and resulted in longer processing times.

Dr. Ramanan’s approach is different in that by analyzing pinpointed sections of the body, the computer has to do less work because these specific points  don’t have an endless amount of possibilities attached to them.

Think of it this way: If I were to hand you a blueprint for a house and it only had a chimney, a door and a window drawn on it, the house you would create would be vastly different from the one I was thinking of. On the other hand, if I were to give you a meticulous blueprint instead detailing every square inch of the house, your finished product would more closely resemble the image in my head. This is the type of work that Dr. Ramanan is doing.

His advancement in Computer Vision has allowed computers to take real objects and convert them into virtual objects, and then use those computer models however necessary.

While direct daily use of this technology is still a few years away, you are probably already using some form of this technology in your everyday life. If you ever find yourself snapping a photo with your iPhone, or even your digital camera, your device will recognize and focus on the faces of individuals in your portrait by forming a square around it. This is a subset of Dr. Ramanan’s work, analyzing the characteristics of a face, recognizing it on the camera and then allowing the camera to focus its lens with respect to that spot.

The technology even extends to gaming, with motion sensor devices such as the Microsoft Kinect using Computer Vision techniques to accurately envision what it is you’re doing, and then transfer that into real-time games. Motion capturing techniques in films could also eventually use this type of technology, instead of all the advanced body suits used in movies like “Avatar.”

But what Dr. Ramanan really wants to do with this technology is use it toward health applications.

“You could put up cameras,” he said, “and use them to monitor progress of patients to see if they need help. It’s nice that it can go towards something altruistic like that.”

In fact, Dr. Ramanan has been working for a while with Dance and Mechanical Engineering professors to develop a game of sorts to help people go through physical therapy. He’s also helped out at places such as Microsoft, the University of Chicago and most recently, Google.

And while his work is still relatively new, he has received multiple awards and honors within the past five years.

“This is my life’s passion,” Dr. Ramanan said. “I see myself doing this for a long time.”