Computer Vision News - May 2018

34 Saturday ISBI DAILY Imaging through scattering is a big one. Objects people have been thinking about this and trying to figure it out forever. It’s a hard problem because it’s very non-linear and you have a lot of degrees of freedom. What’s happening now in computational imaging is we have the computational power to handle gigantic datasets, to handle really massive calculations, which is basically I think what we need for imaging through scattering because it’s such a complicated problem. People have tried so hard to simplify it and with limited results, and now we have a chance to do the more complicated problems potentially. A lot of it is coming from the advances in computing and the advances in just processing giant amounts of data. That’s also true for computational photography, they’re doing a lot of stuff with machine learning and training on these massive image datasets that just weren’t available a while ago. Do you have any thoughts on how academia and industry are interacting today in your field of work? Are there any improvements you would like to make? Computational imaging gets a lot of flak for being around for however long it has been around and not having made enough impact in industry. A lot of people think that it’s mostly just cute tricks that don’t ever make it to real products. It’s starting to make some real products, but a lot of people think it should have been more, given the time and energy people have put into it, so there is some criticism around that. Industry is picking up on this idea that the big tech companies are hiring in this space and trying to find people who are at this intersection between imaging system design and computation. It’s getting better. I do think that it’s true that academia sometimes is working on things that may never be made a product, but it’s valuable when you’re doing something new to just try a bunch of things and see what’s possible, even if you know not all of it is going to be useful. Some of the criticism is warranted and some isn’t, but it’s great that right now industry is starting to care more about research in this space, even if it’s very pre- commercial. What do you think the differences are between your generation of professors and the previous generation? Is there anything that one can learn from the other? I’m super happy with my generation of professors in my field, because we are a very friendly community and Laura Waller I do think that it’s true that academia sometimes is working on things that may never be made a product, but it’s valuable when you’re doing something new to just try a bunch of things and see what’s possible, even if you know not all of it is going to be useful. ” BEST OF