CEVA Introduces Deep Learning in a Chip

8 October, 2015

Ceva Deep Neural Network (CDNN) enables embedded systems to perform deep learning tasks. "Our new technology is a first of its kind in the embedded industry"

“Our new Deep Neural Network framework for the CEVA-XM4 is the first of its kind in the embedded industry”

CEVA

CEVA, Inc. from Herzliya, Israel has revealed a ground breaking software framework to perform neural network deep learning task on its CEVA-XM4 DSP chip. The new Ceva Deep Neural Network (CDNN) enables embedded systems to perform deep learning tasks 3x faster than the leading GPU-based systems while consuming 30x less power and requiring 15x less memory bandwidth.

For example, running a Deep Neural Network (DNN) based pedestrian detection algorithm at 28nm requires less than 30mW for a 1080p 30 frames per second video stream. “Our new Deep Neural Network framework for the CEVA-XM4 is the first of its kind in the embedded industry,” said Eran Briman, vice president of marketing at CEVA.

Key element in the solution is the CEVA Network Generator, a proprietary automated technology that converts a customer’s network structure and weights to a slim, customized network model used in real-time. This enables a faster network model which consumes significantly lower power and memory bandwidth, with less than 1% degradation in accuracy compared to the original network.

Eran Briman, vice president of marketing at CEVA

Phi Algorithm Solutions, from Toronto, Canada, has used CDNN to implement a CNN-based Universal Object Detector algorithm for the CEVA-XM4 DSP. This is now available for application developers and OEMs to run a variety of applications including pedestrian detection and face detection for security, ADAS and other embedded devices based around low-power camera-enabled systems.

The CDNN software framework is supplied as source code, extending the CEVA-XM4’s existing Application Developer Kit (ADK). It works with various networks and structures, such as Caffe, Torch or Theano training frameworks, or proprietary networks. CDNN includes real-time example models for image classification, localization and object recognition. For more information on CDNN: http://launch.ceva-dsp.com/cdnn.

CEVA is a leading licensor of DSP and IP platforms for cellular, multimedia and connectivity. Earlier this week the company announced that its revenues for the third quarter of 2015 are approximately $16.2 million, an all-time high, and above the high end of the Company’s previously revenue guidance of $15 to $16 million. As a result, its NASDAQ stock price rose from $18 to approximately $22. CEVA’s market cap jumped to $446 millions.