Computer Vision News - December 2018

36 Women in Computer Vision Computer Vision News Women in Science interested in engaging in the academic community. This is a way to meet us all at once. The few of us that were there have been keeping in contact, and that’s been really useful. We’re trying to figure out who is in this community and having a list of names. I don’t think we all know each other. It was actually quite random that Adriana Kovashka and I got connected. She knew someone at Columbia who does a lot of video processing related stuff. He happened to know my work. It was really coincidental how we found each other at all. With these coincidences, you find a couple then you find a couple more. It’s really a huge problem. There are computer vision problems. There are human communication problems. There are some ethical issues as well like which kind of images should we be showing people. There are business- related questions. What ethical related questions? In advertising, you associate your product with some image, usually what your customers want. There are many kinds of advertising. In the 1950s, advertising was only about the product. “ We made a microwave. Buy it or don’t buy it! ” You advertised exactly what the product was. Now, we’re in an age where you’re selling a lifestyle. You’re putting a message out there. You want people to believe it. If you have really powerful techniques for that, you have to consider how do we really want to choose to influence people? It’s hard to say this for certain, but I’ve heard a lot of arguments that say that the emerging desktop computer industry in the 80s advertised to boys rather than girls. That had a huge effect. Then they advertised gaming. That really is the first application. That has a huge effect on society. In advertising, you must find your market segment. You can’t advertise to everyone at once. These things have effects. I’m hoping with computer vision, we don’t have to segment markets anymore. We can find out why it’s useful to everyone and find the right customers. That’s the idea in advertising, to find the people that are right for it, and give them the information that they want. What should our magazine Computer Vision News do in that respect? Should we also use visual metaphors? Good question! The first thing you want to think about is who is your audience. You don’t always need a visual metaphor. If you want to convey more of the details in the mechanisms of how an algorithm really works, you want more of a visual explanation. If you wanted to associate something concrete with something more abstract as a way to draw people into the message, maybe use the fact that there are interviews about computer science. And computer scientists! The human aspect which is behind the technology! I think it’s that idea about getting to know someone. That’s one thing about academia: we get to know people’s work and we invite them for talks. I’m kind of surprised, in a good way, to see the publication that you made. I’ve never seen anything like it before in academia. I always enjoy getting to know the person and seeing what motivates them. You never put that in the paper. No one ever says: “ Well, I was walking around one day, and I saw this visual metaphor! Well, that’s kind of cool! How’d they do that? ” We never do that! One more message for our readers? Yes, I’m always looking for interesting students!