Bookmark and Share

NXP Demonstration of Traffic Sign Recognition with Neural Networks

Rafal Malewski, Head of the Graphics Technology Engineering Center at NXP Semiconductors, demonstrates the company's latest embedded vision technologies and products at the May 2017 Embedded Vision Summit. Specifically, Malewski demonstrates traffic sign recognition with neural networks. This embedded solution supports real-time vision and rendering. Its object recognition capabilities employ traditional segmentation along with a convolutional neural network (CNN) engine to classify traffic signs. It utilizes the multi-core graphics architecture of the i.MX 8 SoC, implementing load balancing for the OpenCL/OpenVX segmentation, OpenGL ES rendering and OpenCL CNN, across dual 3D GPUs.