yblaz - Fotolia

Microsoft brings Azure AI to the edge with Azure Percept

Azure Percept includes software for building machine learning models or deploying pre-built models, as well as hardware to deploy the models at the edge.

Microsoft unveiled the public preview of Azure Percept, a platform of hardware and software for bringing Azure AI services to the edge.

The platform, revealed during Microsoft Ignite 2021, includes Azure Percept Studio, software to develop AI models or find and deploy pre-built models, plus a development kit.

No-code edge AI

The kit contains a carrier board with built-in Wi-Fi and Bluetooth connectivity and an NXP iMX8m processor, and an intelligent camera, Azure Percept Vision. Users can also buy Azure Percept Audio, an accessory microphone device that adds speech AI capabilities to the development kit.

The platform, with its no-code, drag-and-drop Azure Percept Studio software and its pre-built hardware devices, is designed to simplify building and deploying AI at the edge.

"I think Azure Percept will make it much, much easier for developers who are not AI experts to build AI edge solutions," said Forrester analyst Mike Gualtieri.

Azure Percept Studio contains sample models for detecting faces, people, vehicles, objects and on-shelf products.

The software is integrated with other Azure services, including Custom Vision, a program within Azure Cognitive Services for training and evaluating vision models, enabling users to quickly train and deploy custom vision models.

Azure Percept comes as advances in hardware and software make it easier to embed AI technology in edge devices.

Azure Percept product image
Microsoft introduces Azure Percept

Better hardware

Artificial intelligence hardware for edge computing has to optimize for power consumption, operating environment, size and, sometimes, heat generation, Gualtieri said.

"Chip manufacturers are increasingly working on inferencing chips that will meet the full combination of edge requirements," he said.

Last December, AWS unveiled edge AI devices designed specifically for industrial and manufacturing clients. The devices, which include IoT sensors to enable predictive maintenance and hardware to add computer vision to existing on-premises cameras, are backed by built-in AWS machine learning services and IoT services.

While AWS and other technology vendors sell hardware with similar capabilities to Azure Percept, the "key advantage of Azure Percept is for the Microsoft developer community," Gualtieri said.

Over the last few years, Microsoft has made strides to provide Azure developers with tools to more quickly and easily use programs within the Azure ecosystem.

"I don't think it will peel away AWS developers to the Azure camp, but it will make Azure developers AI edge developers," Gualtieri said.

Microsoft Ignite 2021 is being held virtually March 2-4.

Next Steps

Microsoft ACS going GA with preview of Teams integration

New Microsoft Azure VMs target generative AI developers

Dig Deeper on AI infrastructure

Business Analytics
Data Management