- Next-generation Cadence Tensilica Fusion DSP Core Expands Capabilities, Aspirations
- CEVA Second-generation Deep Learning Toolset Supports Additional Frameworks and Topologies
- Case Study: How to Implement Deep Learning for Vision on Embedded Processors
- New Synopsys Processor Core Targets Traditional- and Deep Learning-based Embedded Vision
- HSA Foundation Aims for Broader Adoption of Coherent Memory Standard for Heterogeneous Processors
Embedded Vision Alliance Forms to Catalyze Practical Applications of Computer Vision
A new industry association, the Embedded Vision Alliance, is being formed to help embedded system designers harness computer vision in their products. BDTI, which has initiated the partnership, believes that computer vision—extracting meaning from images and video—is poised to proliferate into a wide range of applications in the next few years.
The success of the Microsoft Kinect—which has become the fastest-selling consumer electronics device in history, selling 10 million units in its first few months—is one indication of the opportunities in store for electronic system designers who are able to effectively harness vision capabilities.
Computer vision has been the subject of vigorous academic research for decades, resulting in a deep reservoir of fundamentals and algorithms. But vision algorithms are computationally demanding and, until recently, sufficiently powerful processors were too big, expensive, and power-hungry for use in typical embedded systems. This is beginning to change, though, thanks to new chips that deliver the required processing power at cost and power consumption levels appropriate for high-volume, energy-sensitive systems.
Although computer vision has previously been deployed in embedded systems in selected applications like industrial inspection, surveillance, and automotive safety, the knowledge and skills required to design practical embedded vision systems are not widely available. Often, vision algorithms are developed by researchers who are expert in vision techniques and algorithms but lack knowledge in implementation aspects, while embedded systems are typically designed by implementation experts who don’t have expertise in computer vision.
The Embedded Vision Alliance is a partnership of technology and service companies with expertise in embedded vision applications. These partners seek to catalyze the proliferation of computer vision functionality into a wide range of embedded systems by practical know-how to engineers who are not vision specialists, enabling them to effectively incorporate vision capabilities into their designs. The Alliance will provide content addressing algorithms, processors, tools, and platforms for implementing embedded vision capabilities.
According to Jeff Bier, president of BDTI, “Vision capabilities can add compelling value to an extraordinarily wide range of existing equipment types in markets including medical, automotive, consumer, industrial, and aerospace. For example, we expect to see gesture-based user interfaces added to a variety of systems where current user interfaces are a bottleneck—such as smart phones—and to systems where it is preferable to avoid physical contact between the user and the machine—such as some medical equipment. In addition, we expect to see entirely new kinds of products enabled by the emergence of cost-effective embedded vision capabilities. For example, vision-based safety systems are now being installed in high-end cars.”
The first initiative of the Embedded Vision Alliance will be a web site focused on providing high-quality, practical know-how to system designers in a collaborative setting where both experts and newcomers can gather to share ideas and insights. The site will feature case studies, white papers, blogs, video interviews and demonstrations, and a community forum. The site will launch on May 31, concurrent with the formal debut of the Alliance.
For member companies, the Alliance offers an opportunity to establish a leadership position in embedded vision, while simultaneously helping to accelerate growth of the embedded vision market and attracting the attention of system designers. Companies interested in joining the Embedded Vision Alliance may contact Jeremy Giddings of BDTI at giddings@BDTI.com or +1 (925) 954 1411 for details.