We all know that Microsoft is working on the next version of HoloLens, but so far Microsoft has not officially spoken about it. Today, Microsoft Research team revealed some information about the upcoming HoloLens hardware. The next version of HoloLens will feature the second version of the HPU that will include an AI coprocessor to natively and flexibly implement Deep Neural Networks, able to run resnet18 real-time on the device.
This means that upcoming HoloLens can analyze what the user sees and hears on the device instead of wasting time by sending the data to the cloud for processing. This new AI chip will support a wide variety of layer types and HoloLens team can customise it to their needs. The AI coprocessor can run continuously, off the HoloLens battery, allowing features such as more sophisticated hand tracking with hand-part segmentation (demonstrated below in a video recorded by an attendee) and on-device voice recognition without having to send samples to the cloud.
“This is just one example of the new capabilities we are developing for HoloLens, and is the kind of thing you can do when you have the willingness and capacity to invest for the long term, as Microsoft has done throughout its history. And this is the kind of thinking you need if you’re going to develop mixed reality devices that are themselves intelligent. Mixed reality and artificial intelligence represent the future of computing, and we’re excited to be advancing this frontier,” wrote Marc Pollefeys, Director of Science, HoloLens.
Microsoft HoloLens is made up of specialized components like Holographic Processing Unit that makes light work of processing a large amount of data per second. You can read more about Holographic Processing Unit (HPU) here. At an event in Honolulu today, Harry from MSR team showed an early spin of the second version of the HPU running live code implementing hand segmentation.
On a related note, I can’t wait to hear about the improved optical stack in the next version of HoloLens. I really hope we will see HoloLens with improved field of view.