SAN JOSE, Calif. – An expanding crowd of developers peeked into the future of computer vision at the annual summit of the Embedded Vision Alliance here.
The 33-company alliance gathered well over 300 attendees to see demos and talks on the technologies and the diverse markets it hopes to serve. Demos showed embedded vision applications in automotive, consumer, industrial, medical, robotics, security and other fields.
“Automotive safety is one of the most rapidly maturing markets for embedded vision today, said Jeff Bier, founder of the alliance and president of consulting form Berkeley Design Technology Inc.
Engineers have downloaded more than six million copies of the OpenCV open source library, one of the fundamental software planks of the effort. The code is most notably used in the StreetView feature of Google Maps, but serves a wide range of uses, said Gary Bradski, director of the OpenCV Foundation widely seen as father of the software.
Nvidia is working to accelerate OpenCV on its graphics processors, but the code lacks a hardware accelerator for the ARM architecture. That’s something Bradski hopes changes as momentum builds behind the specification which could hit a version 3.0 this year.
Click on image to enlarge.
Computational photography enables panoramas, better back-lit images and more.
Engineers working on computational photography for mobile systems are turning to OpenCV, Bradski said. Use of the library for that app will be one focus of the Google Summer of Code this year, an open source program for students.
We provide a tour of a handful of the latest demos in the following pages.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.