This year’s Collaborative Robots, Advanced Vision & AI Conference (CRAV.AI) held on October 24-25, 2018 in San Jose, U.S.A. presents the latest developments in robotics and vision and combines them for the first time with Artificial Intelligence (AI). Users, OEMs, manufacturers and integrators from the automation industry will find something of interest, whether they want to implement their first automation system, optimize their existing system or sharpen their understanding of the latest technologies. One of them is Deep Learning within AI, which will be our focus at booth 93 in the exhibitor area.

Our Deep Learning demonstrator will show how FPGA processors can accelerate image processing applications many times over. Trained neural networks of different sizes can be integrated into an application design within our graphical development environment VisualApplets and is combined with image pre-, post- and signal processing. Software developers and application engineers are able to implement such an application design in a short time without hardware programming knowledge. The DL implementation on the FPGA under VisualApplets with the creation of a customized network and image enhancement steps took only two days. The demo application checks defects on reflecting metallic surfaces at a speed of 220 MB/s and a prediction accuracy of 99.4%.

Deep learning high-speed application with frame grabber FPGA for the detection of defects on reflecting metallic surfaces

Contact to our Sales dept. Region Americas
Mike Faulkner
Director of Business Development and Sales
Tel: +1 888 808 3670 Ext 101
Mobile: +1 514 663 7292
E-mail: americas@silicon.software