Kinect is back, and it wants to make the Internet of Things smarter and less reliant on the cloud. Microsoft has taken the wraps off Project Kinect for Azure today at Build 2018, its annual developer event, and if you thought uses for the fiendishly clever depth-sensing camera began and ended with Xbox gaming, think again.
Rather than just a camera peripheral, Project Kinect for Azure will be whole package of sensors. As well as the time-of-flight depth camera, there’ll be onboard compute for standalone processing. It’ll also be small and power-efficient, Microsoft promises.
Why might you want something like that? As Microsoft explains it, the cloud may be ubiquitous but there’ll be times you don’t want to rely on it. The growth of so-called “edge computing” will see increasingly smarter devices develop, able to do their own local processing before feeding what’s important to the cloud.
There are speed benefits there, as well as bandwidth optimizing. We’ve already heard Intel’s argument for why autonomous driving and connected cities will find edge computing essential, for example. Now, Microsoft says there’s the potential for increasing numbers of smart, AI-blessed devices that aren’t reliant on remote processing to do their job.
Project Kinect for Azure builds on not only the original Kinect sensor bar – now a full eight years old – but Microsoft’s more recent hardware like HoloLens. At the same time, Microsoft has a new Speech Devices SDK, which has been developed to improve speech recognition. That means better noise cancellation and far-field microphones, with a range of applications from smart speakers through to drive-thru ordering systems.
It’s not a new topic for Microsoft, of course. Last year, the company revealed Project Brainwave, a deep learning acceleration platform intended for real-time AI. As of BUILD 2018, that’s being launched in preview, complete with full integration with Azure and on the edge.
Why should you care? Although Microsoft’s focus right now is on developers using these new tools and hardware, the promise is smarter, more connected products for you as an end-user. Think smart speakers that are better at hearing you across the room, and faster to respond to your requests; connected devices that have the onboard AI to understand gestures and speech, without the latency and bandwidth requirements of today’s hardware.