Videos

Alexey Rybakov, Senior Director at Luxoft, demonstrates the company's latest embedded vision technologies and products at the September 2016 Embedded Vision Alliance Member Meeting. Specifically, Rybakov demonstrates a new computational photography technique called occlusion (obstruction) removal, running in near-real time on a smartphone platform.

Aaron Behman, Director of Corporate Strategy and Marketing for Embedded Vision at Xilinx, demonstrates the company's latest embedded vision technologies and products at the September 2016 Embedded Vision Alliance Member Meeting. Specifically, Behman demonstrates how the company's SDSoC tool chain can hardware-accelerate embedded vision algorithms.

Paul Karazuba, Director of Marketing for Rambus' Computational Sensing group, demonstrates the company's latest embedded vision technologies and products at the September 2016 Embedded Vision Alliance Member Meeting. Specifically, Karazuba demonstrates the company's LSS lensless smart sensor product.

Steve Kuznicki, Pilot Engineer at MathWorks, demonstrates the company's latest embedded vision technologies and products at the September 2016 Embedded Vision Alliance Member Meeting. Specifically, Kuznicki demonstrates the company's IP core generation workflow, which takes advantage of the HDL Coder automatic code generation support in MATLAB and Simulink, and targets a Xilinx evaluation board.

Bence Varga, Business Development Manager at AdasWorks, demonstrates the company's latest embedded vision technologies and products at the September 2016 Embedded Vision Alliance Member Meeting. Specifically, Varga demonstrates the CNN deep learning-based object recognition portion of the company's self-driving vehicle software stack.

Sergey Federov, Program Manager at Luxoft, demonstrates the company's latest embedded vision technologies and products at the May 2016 Embedded Vision Summit. Specifically, Federov demonstrates a version of the Caffe convolutional neural network framework for deep learning that the company has ported to Android. Object recognition is greatly accelerated by the GPU via OpenCL.

Sergey Federov, Program Manager at Luxoft, demonstrates the company's latest embedded vision technologies and products at the May 2016 Embedded Vision Summit. Specifically, Federov demonstrates the company's Pixerra OpenCL-based, GPU-accelerated real-time video processing library, particularly showcasing HDR and video stabilization functions.

Sergey Federov, Program Manager at Luxoft, demonstrates the company's latest embedded vision technologies and products at the May 2016 Embedded Vision Summit. Specifically, Federov demonstrates a platform-optimized version of the OpenCV library histogram calculation function, which runs approximately 4x faster than the non-optimized function version.

Sergey Federov, Program Manager at Luxoft, demonstrates the company's latest embedded vision technologies and products at the May 2016 Embedded Vision Summit. Specifically, Federov demonstrates the company's OpenCL-based Multi Neural Network System, in this particular case used in an offline fashion for object recognition (via R-CNN) and tagging purposes.