Industry standards are critical to enabling widespread adoption of any new technology. With respect to embedded vision, they include such topics as APIs (application programming interfaces) and algorithm libraries. For example, Khronos' new OpenVX standard enables cross-platform acceleration of computer vision algorithms. Khronos' OpenCL framework is often used for writing computer vision programs that execute on heterogeneous processing platforms. And the popular open-source OpenCV library of computer vision functions, while not a formal standard, is a de facto standard for many developers.
All of these subjects will be showcased at the Embedded Vision Summit:
- Gary Bradski, head of the OpenCV Foundation and one of the most recognizable names in computer vision, will present "The OpenCV Open Source Computer Vision Library: Latest Developments." Bradski will provide an insider's perspective on the new OpenCV 3.0 and how developers can utilize it to maximum advantage for vision research, prototyping, and product development.
- Deshanand Singh, Director of Software Engineering at Altera, will discuss "Efficient Implementation of Convolutional Neural Networks using OpenCL on FPGAs." Singh will give a detailed explanation of how CNN algorithms can be expressed in OpenCL and compiled directly to FPGA hardware.
- Mostafa Hagog, Architect at Intel, will talk about "Developing High-Performance, Energy-Efficient Vision Solutions Using Open-Standard Libraries and APIs." Using real-world examples, Hagog will explore how open standard libraries and APIs can speed application development while often also improving performance and energy efficiency.
- Yury Gorbachev, Principal Engineer at itseez, will speak about "OpenCV for Embedded: Lessons Learned." Gorbachev's talk will address topics such as cross-platform development, performance profiling, benchmarking, and automated regression testing.
- And Neil Trevett, President of the Khronos Group, will present "Navigating the Vision API Jungle: Which API Should You Use and Why?" Trevett will map the landscape of APIs for vision software development, highlighting where they overlap and complement each other, and will preview some of the latest developments.
In addition, two workshops taking place in conjunction with the Summit cover various industry standards:
- "Implementing Computer Vision: A Technical Introduction," developed by BDTI in collaboration with the Khronos Group, is a workshop on algorithms, processors, sensors, tools, libraries, and standards for practical computer vision systems and applications. The workshop gives special attention to OpenCV, OpenCL and OpenVX. It takes place on May 11th from 9 AM to 5 PM, and the $299 registration fee includes lunch.
- Altera's "SDK for OpenCL for Embedded Vision Applications on FPGAs" workshop will explore how the Altera SDK for OpenCL abstracts away the traditional FPGA hardware development flow and provides a software-familiar flow that allows the FPGA to be prototyped in minutes. It runs from 8:30 AM – 12:30 PM on May 13th and registration costs $150.
Head to the Embedded Vision Summit section of the Alliance website for in-depth information on these and other technical presentations and workshops, as well as information on the Summit keynotes 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; accompanying half- and full-day workshops are offered on May 11 and 13. Last year's Summit sold out, so don't delay, register today!
