CEVA, Inc. the leading licensor of signal processing IP for smarter, connected devices, has announced that the CEVA Deep Neural Network (CDNN2) software framework has been honored with the “2017 Most Innovative Product” award from OpenSystems Media’s Embedded Computing Design.
“We are delighted to be a recipient of this year’s “Most Innovative Product” award from Embedded Computing Design,” said Ilan Yona, vice president and general manager of the Vision Business Unit at CEVA. “There is tremendous industry momentum behind artificial intelligence and deep learning technologies and our CDNN2 software framework enables the rapid deployment of advanced neural networks in any CEVA-XM powered device. This prestigious award is another endorsement of our complete platform strategy to addressing the mass-market AI challenge.”
The criteria for the product choices boiled down to a few key areas: technology innovation, relative performance, and potential business/market impact. CDNN2 received high marks from judges in both Design Excellence and Market Disruption categories. The judges came from Embedded Computing Design’s esteemed Advisory Board and the Content/Editorial Team.
“It is our pleasure to recognize CEVA in our 2017 Most Innovative Product Awards for their CDNN2 software framework,” said Patrick Hopper, President of OpenSystems Media. “A key feature of the toolkit is its ability to eliminate months of development time for engineers developing intelligent products, which in turn serves to accelerate the use of artificial intelligence in our everyday lives.”
CDNN2 is a comprehensive toolkit that simplifies the development and deployment of deep learning systems for mass-market embedded devices. Tailored and optimized for the CEVA-XM family of imaging and vision DSPs, the CDNN toolkit includes the CEVA network generator, the CDNN real-time software framework, and a CNN hardware accelerator that works together to deliver superior performance while ensuring flexibility with the constantly evolving requirements of machine learning. Using a simple push-button mechanism, CDNN2 converts any pre-trained neural network to a network optimized for CEVA-XM based embedded systems while maintaining 99 percent accuracy and allowing it to run on a low power DSP; running CDNN2 on CEVA-XM DSPs improves power efficiency by up to 25x, with up to 4x faster processing than GPU- or CPU-based systems.
Source: CEVA, Inc.
