Renesas Electronics Corporation, a provider of advanced semiconductor solutions, has expanded its RZ/V series of AI-capable microprocessors (MPUs) with a new device that enables AI processing of image data from multiple cameras, offering a new level highly accurate imaging. recognition for vision AI applications.
Equipped with two 64-bit Arm Cortex-A53 cores, the new device is capable of delivering high computational performance with a maximum operating frequency of 1GHz. The RZ/V2MA features a proprietary low-power dynamically reconfigurable processor accelerator (DRP-AI), which can process vision AI with class-leading performance of 1 TOPS/W (tera operations per second, per watt).
The RZ/V2MA device offers high-speed interfaces such as Ethernet, USB and PCI Express, allowing input of images from multiple external cameras. In addition to the DRP-AI accelerator, RZ/V2MA includes an OpenCV accelerator that enables rules-based simultaneous image processing.
These capabilities provide highly accurate image recognition capabilities for machine vision products such as AI-powered gateways, video servers, security gates, POS terminals, and robotic arms.
“One of the challenges for embedded system developers who want to implement machine learning is keeping up with the latest AI models that are constantly evolving,” said Shigeki Kato, vice president of Renesas' enterprise infrastructure business division. “With the new DRP-AI TVM tool, we give designers the option to expand AI frameworks and AI models that can be converted into executable formats, enabling them to bring the latest image recognition capabilities to embedded devices using new models of AI.”
The new RZ/V2MA offers a complete set of development tools to assist in the design of vision AI systems. In addition to the existing DRP-AI translator, the new device adds DRP-AI TVM (Note 1), which is based on the open source deep learning compiler Apache TVM technology (Note 2).
Although DRP AI Translator is designed to convert AI models into DRP-AI executables, the DRP-AI TVM compiler allows the DRP-AI accelerator to work in conjunction with the CPU, allowing DRP-AI to convert and generate more AI models than ever before. As a first phase, Renesas supports ONNX and PyTorch AI models and plans to support Tensorflow in the future.
“Renesas’ RZ/V series is ideal for embedded devices as it does not need fans or heatsinks due to its extremely low power consumption and low heat capacity when running AI,” said Chiharu Nakabayashi, President of amnimo Inc ., a provider of IoT and AI-based services and a wholly owned subsidiary of Yokogawa Electric Corporation. “With these devices, we are confident that we can develop powerful imaging AI gateways that can be installed anywhere.”
