IP Track Invited

27.3 - eFPGA for Neural Network based Image Recognition

1:30 PM–3:00 PM Jun 26, 2018 (America - Los Angeles)

Room 2008

This talk explains how to build a monolithic device performing Image Recognition thanks to Menta embedded FPGA IP. The device associates sensors and advanced neuronal processing combined with eFPGA for both data pre-processing and local decision making. Menta deliver single core eFPGA IPs that are 100% standard cells based for best-in class integration and verification within customers EDA design flow. The eFPGA IP performances and flexibility allow high bandwidth data pre-processing with the exact amount of resources required, targeting the right process node, selected by the customer. Menta eFPGA IP architecture and programming software capabilities are detailed. Learn what parameters are key in choosing an embedded FPGA solution, thanks to Menta 10+ years of experience in eFPGA IP.