As Artificial Intelligence continues its relentless evolution, one of the major challenges facing the technology is how to successfully integrate it into the small, low power devices we use in our everyday lives. The extremely-deep neural networks underlying AI require high-performance, yet power-efficient embedded platforms to handle the processing load. CEVA offers a scalable imaging and vision processor, combined with its neural network software framework and unique “push button” network converter which converts pre-trained networks to real-time optimized networks for embedded devices. This solution enables complete flexibility with regards to performance and target application, while significantly reducing time-to-market. In this presentation, we will introduce the latest CEVA vision processor and demonstrate how to implement an embedded neural network using our tools in a matter of minutes, rather than months.
CEVA will be demonstrating CNN-based demos powered by CEVA’s latest XM4 vision processor. In addition we’ll be discussing CDNN2 software framework, which is a quick and automated way to convert, optimize and execute a pre-trained networks on embedded platforms and a great time-to-market.
Please join CEVA team at the Linley Processor Conference that will be held on September 27 – 28, 2016 at the Hyatt Regency Hotel, Santa Clara, CA.
You are invited to sign an appointment with me.
Liran Bar
Ceva