THE TAKEO KANADE ENDOWED PROFESSORSHIP
In August, Carnegie Mellon University announced the creation of the Takeo Kanade Professorship, in honor of one of the genuine trailblazers in the field of computer vision.
CHRIS QUIRK
The professorship is the gift of a group of longtime friends and former students — Jing Xiao and Yu Li, Hongwen Kang, Yan Li and Chenyu Wu, Hua Zhong and Min Luo, Lie Gu and Yi Zhou, Mei Han and Wei Hua, Yanghai Tsin and Hongming Jin and Hang Su — who came together to find a meaningful way to celebrate Takeo Kanade’s achievements. Their joint lead-gift for the Kanade Professorship recognizes the importance of Kanade’s career and how he changed the course of their lives as a teacher and mentor.
Takeo Kanade is the U. A. and Helen Whitaker University Professor of Computer Science and Robotics. Over the course of his career, Kanade invented, worked on or published some of the most important creations and theories of computer vision, and also developed unique robotic devices. “Takeo is one of a few early pioneers in the fields of computer vision and robotics. Over the past five decades, his foundational contributions have had a profound impact on further reaching fields of study and he has made seminal contributions in areas ranging from computer vision and autonomous vehicles to medical robotics,” said Martial Hebert, dean of the School of Computer Science. “He shaped the field of computer vision from its infancy in the 1970s to its current level of exponential growth. Within CMU, Takeo has been instrumental in transforming the Robotics Institute from a research center into a full-fledged academic department with graduate and undergraduate programs that are unique.”
“Takeo is one of a few early pioneers in the fields of computer vision and robotics.”
Born in Hyogo, Japan, Kanade earned his doctorate in electrical engineering from the Kyoto Institute, and then taught there for seven years. In 1980, he began his more than 40-year tenure at Carnegie Mellon. Kanade has won numerous awards for his work, including the prestigious Kyoto Prize for Advanced Technology, which recognizes those who have made significant contributions to the betterment of humanity. Kanade has also won the Bower Award and Prize for Achievement in Science from the Franklin Institute in 2008, and both the Okawa Prize and the ACM-AAAI Allen Newell Award in 2007.
In the 1970s, Kanade created one the first facial recognition systems, using a library of facial images he assembled that was likely the largest database of its kind at the time. In the 1980s, Kanade led a team of Carnegie Mellon researchers that developed NavLab, one of the earliest self-driving vehicles. Two Carnegie Mellon researchers took the vehicle 3,000 miles, from Pittsburgh to San Diego. NavLab drove more than 98% of the trip autonomously.
“He shaped the field of computer vision from its infancy in the 1970s to its current level of exponential growth.”
Kanade’s most famous invention is EyeVision, which he designed and built for CBS to use during Super Bowl XXXV in 2001. For the broadcast, Kanade and his team mounted an array of cameras high up on Raymond James Stadium in Tampa, and designed a computer processing system that synthesized the inputs from the cameras into a smooth, flowing image that rotated the angle of the view of the play. Kanade appeared at halftime to explain the technology, and he may well be the only professor to appear on a Super Bowl broadcast. “Before EyeVision, when I got on a plane and talked to the person next to me during the flight, I’d tell them I was a professor at Carnegie Mellon working on robotics, and that was it,” said Kanade. “But when I told people I built EyeVision, they were very interested to hear about it. EyeVision was a ticket to wonderful conversations with anybody in the world.”
In 1981 Kanade discovered, along with his then-student Bruce Lucas, an algorithm that provided a new way to track objects visually, the Lucas-Kanade Method. “Bruce said we should publish it, but I told him no, you can’t write a paper on this because it’s based on the Taylor Expansion, which is 300-year-old math,” Kanade said. Lucas placed the paper in the Proceedings of the 7th International Joint Conference on Artificial Intelligence. The Lucas-Kanade Method became a formative tool in the field, and the paper has garnered more than 18,000 citations. “It just goes to show you,” said Kanade, laughing, “that if your professor tells you not to publish something, you might want to publish it anyway.”
“EyeVision was a ticket to wonderful conversations with anybody in the world.”
At 78, Kanade is still a dynamo of energy for his work, and remains unceasingly curious. In a recent interview that went almost three full hours and ended well after midnight, Kanade recounted his achievements with his characteristic modest humor; expounded on the history of computer vision, epistemology and the future of technology; and, described future projects. He is now devising a multi-camera technology using mobile phones so remote viewers could tour a spectacular site virtually, an idea he had when visiting the Taj Mahal. “It’s a beautiful, unimaginable place. Everybody should visit, but it’s not easy to get to,” said Kanade.
The Takeo Kanade Professorship will be a legacy not just to Kanade’s remarkable career, but to the spirit and legacy of innovation that Kanade and his colleagues have created at the university. “I don’t think it’s an overstatement to say that the Carnegie Mellon Computer Vision group, for the last 40 years, has been the most advanced and influential,” said Kanade. “We covered everything, from the theoretical to software to technology and hardware. Our versatility is probably our real forte, and I’m proud of that.” ■