Embedded Vision Insights: April 23, 2015 Edition

In this edition of Embedded Vision Insights:

3D Vision and Processor Options
Enabling Capabilities of Low-Cost 3D Vision
Learning OpenCV
Embedded Vision in the News

LETTER FROM THE EDITOR

Dear Colleague,

Some believe that 3D vision is a game-changer, enabling robust, practical solutions to problems that are difficult-to-impossible to solve with conventional 2D vision. 3D enables easier discrimination between objects and their background, for example. It can also enable more reliable and more precise gesture interfaces. And it helps systems understand where they are located in relation to other objects.

This year's Embedded Vision Summit features several presentations addressing 3D vision:

The keynote "Bringing Computer Vision to the Consumer," from Mike Aldred, Electronics Lead at Dyson, will highlight the company's 360 Eye robot vacuum cleaner, which uses 3D vision to map a room and determine its position in it. Aldred will also discuss the challenges of bridging between academia and business, and how to use a diverse team to take an idea from the lab into real homes.
"3D from 2D: Theory, Implementation and Applications of Structure from Motion," a technical presentation by Marco Jacobs of videantis, will explore the theory and efficient implementation of the structure from motion technique, which extracts depth information using only a single, moving 2D camera.
And Ken Lee, president of VanGogh Imaging will present "Bringing New Capabilities to Users and Industries with Mobile 3D Vision." Lee will explore how diverse applications benefit from the ability to separate a scene into discrete objects and then recognize and analyze them reliably.

Another critical topic that will be addressed at the Summit is processor choices for vision applications. Delivering the processing performance required by computer vision applications (typically tens of billions of operations per second) with cost and power consumption appropriate for mass-market products is a tough challenge. Just in the past few months, several new vision-specific co-processors, offered as licensable cores for incorporation into chips, have been announced, joining multiple previously introduced vision cores. In addition, many vision system developers use other types of parallel co-processors, such as GPUs, DSPs and FPGAs.

This wealth of vision processor options is great news for chip and system designers, because a range of application-optimized processor choices makes it more likely that you'll be able to find a processor that fits your specific needs. At the same time, the large number of diverse processor options, and the rapid pace of new options being introduced, can make it difficult to make the best selection decision. Multiple Embedded Vision Summit presentations and other events strive to simplify this task:

Jeff Bier, Embedded Vision Alliance founder and President of BDTI, will present the talk "Choosing a Processor for Embedded Vision: Options and Trends." Bier will highlight strengths and weaknesses of different processor types, as well as illuminate important trends in processors and associated development tools.
In "Understanding the Role of Integrated GPUs in Vision Applications," Roberto Mijat, Visual Computing Marketing Manager at ARM, will explore when it makes sense to utilize the GPU as a coprocessor for computer vision algorithms, what to expect from the GPU, and other key considerations.
Chris Rowen, Fellow at Cadence Design Systems, will speak about "Designing and Selecting Instruction Sets for Vision." Rowen's talk will serve as a step-by-step tutorial on how to dissect vision applications, extract key requirements, and determine processor instruction set and memory organization priorities.
And two of the workshops taking place in conjunction with the Summit, both on Monday afternoon, May 11 and free of charge, will also focus on processor topics. "Enabling Computer Vision on ARM" will encompass presentations from multiple computer vision experts, who will share their experiences working with ARM-based systems across a variety of real use cases. And "Using the DesignWare Embedded Vision Processor for Low-Power, Low-Cost Video Surveillance and Object Detection Applications" will educate you on the features of Synopsys’ new embedded vision processor core, in the context of several detailed case studies.

Head to the Alliance website for in-depth information on these and other presentations and workshops, as well as information on the other Summit keynote, from Baidu's Dr. Ren Wu, and the technology showcase. The Embedded Vision Summit, an educational forum for product creators interested in incorporating visual intelligence into electronic systems and software, takes place on May 12, 2015 at the Santa Clara (California) Convention Center; accompanyinghalf- and full-day workshops are offered on May 11 and 13. Last year's Summit sold out, so don't delay, register today!

While you're on the Alliance website, make sure to check out all the other great content recently published there, such as an interview with Satya Mallick, a computer vision and deep learning entrepreneur and enthusiast. And to further enrich your technical expertise in vision-capable applications, consider attending the May 6-7 Internet of Things Developers Conference, also taking place in Santa Clara, California. Thanks as always for your support of the Embedded Vision Alliance. Please don't hesitate to let me know how the Alliance can better serve your needs.

Brian Dipert
Editor-In-Chief, Embedded Vision Alliance

If you're building AI or vision-enabled products, you've come to the right place.

Embedded Vision Insights: April 23, 2015 Edition

Pages

Topics

Contact

Address

Phone