Research Lives: Dr Suzanne Little, Dublin City University lecturer and researcher in the Insight Centre for Data Analytics

We ‘teach’ computers how to look for things in pictures

You teach computers how to see – tell us more Computers are good at a lot of things, but seeing isn't one of them. When we humans look at a picture we can make sense of it quickly, but it is impractical for humans to sift through millions of images or hundreds of hours of video.

So we develop software to “teach” computers how to look for things in pictures, then they can do the hard labour of pulling that information out of mountains of images.

When might this be useful? We work on a few different situations to help people get more useful information out of video footage. One is the Smart Stadium project between DCU, Arizona State University, Croke Park and Intel. In that case, computers being able to track crowds from camera images around the stadium [could] help manage those crowds more easily as they move around. If something happens, say someone falls or there is a surge, it is really hard for a human to find that, but if a computer knows what to look for, it can find it and alert a human. We also use cameras and imaging software to observe the grass on the pitch and we can analyse footage from games to help players and coaches figure out how they can up their game.

But what about privacy? That's a fundamental issue and it is something we factor into the design of our projects. We have to adhere to EU regulations about privacy and data protection, and we also design software in a way that it doesn't matter when the faces are blurred. For example, if you just want to know how crowded an area is, we don't need to see people's faces, so can we blur those faces quickly in the process and render the information as data where people are not identifiable. Sometimes people ask me if all this tracking is a bit "Big Brother-ish", but I think it is good to talk about the ethics and figure out how we can get the information we need without compromising people's privacy.

READ MORE

Have you always been interested in science? When I was about six years old, I was playing with a drinking straw, putting my finger over the top and I wondered why liquid stayed in the bottom of the straw when I lifted it out of the drink. I made a poster about the surface tension that caused it for a science competition in Queensland, Australia, where I'm from, and I won. That put me on track for a lot of science programmes and visits to labs when I was a teenager.

Do you pass on the love of science and tech now? Yes, I work with Girls Hack Ireland. We run "hackathons" to encourage teenage girls to learn about computers and to feel comfortable working on the technology behind websites and games.

What gets you out of bed in the morning to do research? I think it is the problem-solving aspect. It can be small problems, like needing more computing power to do this algorithm, so is there a clever way I can change my algorithm to overcome that? Or it might be thinking forward to anticipate what will happen when an algorithm runs and how to avoid problems.

What do you do when you are not being a scientist? Is a scientist ever not being a scientist? When I'm not doing research I like to travel, and I enjoy music. I sing in the Goethe Institute choir, which is great because it keeps me away from the computer screen.