Embedded vision processor adds AI for 3D imaging chips

Jon Lawson

Synopsys has added machine learning (ML) and artificial intelligence (AI) tools to its EV62 Embedded Vision Processor IP to allow the development of a high performance 3D imaging system-on-chip design.

The NU4000 from Intuitive is using the tightly integrated vector DSPs and convolutional neural network (CNN) engine in the EV62 to implement the ML algorithms for high definition 4K cameras. The CNN engine operates in parallel with the vector DSPs, enabling more accurate object detection, image classification and semantic segmentation in the chip with a fraction of the power consumption of competing vision processor IP solutions. 

The EV6x processors, based on the ARC configurable instruction set, are supported by the MetaWare EV Development Toolkit, a comprehensive set of tools and software. These tools distribute the NU4000's computations between the vision processor core and CNN resources to support new and emerging neural network algorithms as well as customer-specific CNN layers. 

They support the OpenVX, OpenCV and OpenCL embedded vision standards as well as a new CNN mapping tool that automatically converts neural networks trained using popular frameworks like Caffe and Tensorflow into an executable for the CNN engine.

"To provide our customers with the most advanced, high-performance vision processing capabilities for their demanding computer vision and machine learning applications, we require vision processing IP with the best mix of performance and power consumption with the flexibility to support current and future CNN graphs," said Dor Zepeniuk, vice president of R&D at Inuitive. "The competitive power, performance and area advantages of the DesignWare EV62 processor with its fast CNN engine were critical to achieving the real-time processing capabilities of our NU4000 3D imaging and vision processor SoC."

The EV6x processors have a heterogeneous multicore architecture with up to four vector DSPs with performance up to 4.5 TeraMACs/sec in today’s mainstream 16nm FinFET chip making process technologies.

"Artificial intelligence is enabling intelligent, real-time identification and recognition at high precision for applications such as ADAS, video surveillance and virtual/augmented reality," said John Koeter, vice president of marketing for IP at Synopsys. "By using Synopsys' EV6x Vision Processor IP, companies like Inuitive can create SoC designs that execute deep learning algorithms at TeraMAC-per-second speeds with the lowest power consumption and highest accuracy."