Skip to main content

Bing Vision technology gets explained

Get Microsoft Silverlight

Over at iStartedSomething, Long Zheng posted a nice little clip from Microsoft Research dealing with 2D object recognition, which for those who have been following, is instantiated in Windows Phone "Mango" as Bing Vision (see demo). The tool of course allows to you scan various 2D objects like CDs, DVDs, posters, barcodes, etc. which can then access a database for pricing and product info.

In this brief segment, Rick Szeliski from Microsoft Research’s Interactive Visual Media group goes into detail of how this works. While it's some heady stuff, most of you should be able to watch the ~4 minute segment and get an understanding of how this stuff happens and what the current limitations are--specifically 3D objects and "generics" likes recognizing your pet or people.

While Google has had Goggles for awhile and all of this is based on pre-existing tech, it's neat to see it explained. Plus it doesn't take the magic away when you actually use Bing Vision on your phone (as anyone who has demoed this to a non-techy can attest too). Pretty amazing work when you think about it.

Read moe at iStartedSomething

Daniel Rubino is the Executive Editor of Windows Central, head reviewer, podcast co-host, and analyst. He has been covering Microsoft here since 2007, back when this site was called WMExperts (and later Windows Phone Central). His interests include Windows, Microsoft Surface, laptops, next-gen computing, and arguing with people on the internet.