Michael Tusch, founder and CEO of Apical, presents the "Harnessing Hardware Accelerators to Move from Algorithms to Embedded Vision" tutorial within the "Using Processors for Embedded Vision" technical session at the September 2012 Embedded Vision Summit. The presentation focuses on the programming challenges associated with implementing vision algorithms efficiently on embedded devices with a degree of hardware parallelization. Tusch illustrates the two fundamental issues of data flow and memory access with simple code examples.

Eric Gregori, senior software engineer and embedded vision specialist at BDTI, presents the "Finding Objects Using Canny Edge Detection" tutorial within the "Using Vision Algorithms" technical session at the April 2013 Embedded Vision Summit. This presentation dives into the Canny edge detection algorithm; how it works and how to use it. Topics include a Canny algorithm walk through, connecting edges with contours, and classifying contours with mathematical models. Finally, the topics learned are applied to describe a system for counting dots on dice.

CogniMem’s Chris McCormick, application engineer, demonstrates how the addition of general-purpose and scalable pattern recognition can be used to bring enhanced gesture control to the Microsoft Kinect. Envisioned applications include augmenting or eliminating the TV remote control, using American Sign Language for direct text translation, and expanding the game-playing experience. To process even more gestures in parallel, simply connect together multiple CM1K neural network processors, in order to expand the knowledge base without impacting recognition performance.

Frank Brill, Vision Software Manager at NVIDIA, represents Khronos in presenting on the OpenVX vision processing API at the April 2013 Embedded Vision Alliance Member Meeting. For a PDF download of Brill's talk, please see the link below.

by Vineet Aggarwal
Embedded Systems Group Manager
National Instruments

Markus Wloka, R&D Director for System-Level Solutions at Synopsys, presents the "Lessons Learned: FPGA Prototyping of a Processor-Based Embedded Vision Application" tutorial within the "Developing Vision Software, Accelerators and Systems" technical session at the April 2013 Embedded Vision Summit. This presentation covers the steps of building a programmable vision system and highlights the importance of having software-driven tests and using a chain of proven tools and components.

José Alvarez, Engineering Director for Video Technology at Xilinx, presents the "Challenges and Opportunities in Accelerating OpenCV" tutorial within the "Developing Vision Software, Accelerators and Systems" technical session at the April 2013 Embedded Vision Summit. Computer vision continues to see increased adoption across a wide variety of applications. The industry's trend towards the incorporation of Smarter Vision capabilities is driving the need for ever more intelligent algorithms that are capable of image understanding.