Share this page
Share this page E-mail this page Print this page RSS feeds
Home > About Microsoft Research > Putting Innovation to Work in Microsoft Products
Putting Innovation to Work in Microsoft Products

(vi) Products

Transforming technological advances into software that blends naturally into people’s everyday lives is not a simple task. Even what sounds like a simple improvement to consumer products — such as developing an algorithm that removes red-eye in photographs or helps people write clearer sentences — can involve complexities that require just as much collaboration as the research itself.

 

“It’s all about exchanging ideas and solving problems. The best technology transfer occurs when the problems can meet the solutions.”
 

Philip Fawcett, technology transfer program manager, Microsoft Research

   

A team of self-styled “technology transfer agents” helps bring together the research and product-producing sides of Microsoft. They spend their days investigating research projects at Microsoft Research, meeting with Microsoft product teams and helping spot places where the research might come in handy. Another group of program managers seeks external uses for valuable inventions that provide opportunities for technological advances. Microsoft Research also holds annual technology festivals on the Redmond campus. The events are similar to science fairs, allowing researchers to demonstrate and display their projects for other Microsoft employees.

The following examples demonstrate the broad variety of contributions that Microsoft Research has made to Microsoft products:

pictureTablet PC. In conjunction with the product development team, contributions include:

  • On-screen inking technology. Methods for improving computer recognition of written Asian characters.

Microsoft Digital Image Pro.

Contributions include:

  • An editing technique that enables people to fix common problems with any photograph, such as red-eye that occurs when the flash hits the pupil in the wrong way.
  • A quick, effective way to fix the brightness and contrast in a photo.
  • Algorithms to make cut-and-paste operations easy.
  • Smart Erase, a photo-editing tool that enables users to remove objects from a picture.
 

“Microsoft Research isn’t about one-time technology transfer. My product team has an ongoing relationship with Microsoft Research. This allows us to find a balance between ideas and opportunities that are purely born out of the imagination and talents of the researchers with the pragmatic engineering that our team has to do.”
 

Marc Olson, group product manager, Microsoft Office Outlook

   

Xbox. Contributions include:

  • Artificial intelligence technology that discerns a player’s level of expertise and adjusts the game to an appropriate level of difficulty, and pairs players within the Xbox Live® online gaming service with similar connection speeds and gaming experience.
  • Graphics technology that allows Xbox developers to offer highly realistic images and model complex images, such as animal fur, wood grain, and blades of grass.
  • Audio compression and optimization technology for more realistic sound effects and faster gaming.

pictureSQL Server. Data-mining algorithms developed by Microsoft Research enable businesses to:

  • Find patterns in massive data sets and integrate the information with the server’s existing inquiry API.
  • Uncover patterns and analyze data stored in the SQL Server database.
  • Classify customers by products bought and incorporate the information into the existing database.
  • Customize data further by adding a large repertoire of their own algorithms and using techniques for machine learning and making statistical inferences.

pictureWindows Media Player. In conjunction with the Microsoft Windows Digital Media Division, contributions include:

  • Architecture with coding and algorithms to compress a Windows Media audio file so that it uses 10 to 20 times less space than the original CD and one-half the space of an MP3 file.
  • A protection feature to manage file distribution.
  • The ability to recover meta-tags on music so that users can identify songs on Internet radio or customized CDs that are not labeled. By analyzing the pattern of the music on each track and matching it with the pattern of the song in Microsoft’s database, the system also corrects any songs that are mislabeled and identifies duplicate recordings.