Posted by
(opens in new tab)In early June, Jamie Shotton (opens in new tab) received a most welcome email from Ramin Zabih, professor of computer science at Cornell Tech and chair of the IEEE Pattern Analysis and Machine Intelligence (PAMI) Technical Committee.
Zabih had the responsibility to let a gifted young individual know that he was the 2014 recipient of the PAMI Young Researcher Award (opens in new tab). That was the message Shotton received.
on-demand event
Microsoft Research Forum Episode 4
Learn about the latest multimodal AI models, advanced benchmarks for AI evaluation and model self-improvement, and an entirely new kind of computer for AI inference and hard optimization.
“I knew I’d been nominated,” he says, “but actually hearing I’d won was rather exciting! I was beaming ear-to-ear for the rest of the day.”
The award recognizes a young researcher for distinguished research contributions to computer vision, and it goes to a researcher within seven years of Ph.D. completion. This year, the PAMI Technical Committee honored two people, Shotton and Derek Hoiem (opens in new tab), an assistant computer-science professor at the University of Illinois at Urbana-Champaign.
“I also found out that I’m sharing the award with my friend and academic colleague, Derek Hoiem,” Shotton says. “I have already greatly admired his work, and I can’t think of anyone I’d rather be receiving this award with.”
Shotton, 33, earned a Ph.D. in computer vision and visual object recognition from the University of Cambridge before joining Microsoft Research Cambridge (opens in new tab) in 2008. His research interests since then have extended in numerous directions, including human pose and shape estimation, object recognition, machine learning, gesture recognition, 3-D reconstruction, and medical imaging.
If the name sounds familiar, it’s likely for his contributions in human pose estimation to Kinect (opens in new tab). That product, which set a Guinness World Record as the fastest-selling consumer-electronics device, included a machine-learning contribution that was recognized by the Royal Academy of Engineering in June 2011. That august body presented the MacRobert Award (opens in new tab) for innovation to Shotton and his Microsoft Research colleagues Andrew Blake (opens in new tab), Andrew Fitzgibbon (opens in new tab), Toby Sharp (opens in new tab), and Mat Cook.
The paper that resulted from that collaboration with Microsoft's Interactive Entertainment Business group, Real-time Human Pose Recognition in Parts from Single Depth Images (opens in new tab), for which Shotton was the lead co-author, won the Best Paper Award during that year’s Computer Vision and Pattern Recognition (CVPR) conference. Demonstrating Shotton’s continued commitment to the field, he will be delivering two oral presentations and one poster during this year’s conference, being held June 23-28 in Columbus, Ohio:
- Filter Forests for Learning Data-Dependent Convolutional Kernels (opens in new tab), by Sean Ryan Fanello, Cem Keskin, Pushmeet Kohli (opens in new tab), Shahram Izadi (opens in new tab), Shotton, Antonio Criminisi (opens in new tab), Ugo Pattacini, and Tim Paek (opens in new tab).
- Multi-Output Learning for Camera Relocalization (opens in new tab), by Abner Guzman-Rivera, Kohli, Ben Glocker, Shotton, Sharp, Fitzgibbon, and Izadi.
- User-Specific Hand Modeling from Monocular Depth (opens in new tab), a poster explaining work by Jonathan Taylor, Richard Stebbing, Varun Ramakrishna, Keskin, Shotton, Izadi, Aaron Hertzmann, and Fitzgibbon.
Those papers, part of a strong and broad Microsoft Research contribution to CVPR 2014, also represent the growing influence Shotton is having on his field.
“It’s been an amazing ride,” he says, “far exceeding my wildest expectations.
“I have many, many people to thank for that. I vividly remember my first internship, in 2003, at Microsoft Research Cambridge, and while I thought I’d love to work there full time, I never dreamed I’d actually be able to make it as a researcher.”
Now, with PAMI Young Researcher Award credentials in tow—“Feels good to still be classed as ‘young,’” he says—Shotton seems poised to take the next step in what seems certain to be a fruitful career.
“I want to continue working with my fantastic colleagues on great core vision, graphics, and machine-learning research,” he says. “But further, I think we at Microsoft Research are now uniquely placed to go after bigger and bolder challenges and build things that have real impact in the world.
“I want to be part of that, building systems that can really understand what they’re looking at, and making computers that feel completely natural to use—and have a lot of fun along the way.”