Microsoft’s experimental Garage team today rolled out Snip Insights that lets users draw information from screenshots taken on Windows 10 PCs.
Snip Insights allows people to do things like take a screenshot of a dress and then find out where it can be purchased, or pull text from images of documents or billboards.
Snip Insights can also recognize things like landmarks, famous people, and locations. Computer vision that powers Snip Insights is drawn from Azure Cognitive Services.
The open source tool, which can be downloaded on GitHub, was made by a group of Microsoft Garage interns, according to a company blog post. Snip Insights draws on Windows Snipping Tool, a retired Garage project designed to help users stylize and share screenshots.
The news comes a day after Google revealed new features for its Lens visual search tool. In addition to news that Lens will be added directly to the camera apps of high-end smartphones like Google’s Pixel and LG G7, on Tuesday Google announced Lens will soon be able to identify fashion styles and perform real-time analysis of objects placed in front of a camera.
Both Snip Insights and Lens compete with visual search products from other companies, such as Amazon and Pinterest, which also has a computer vision-powered service named Lens that can identify styles and other insights from pinned photos.
Each of the aforementioned services is available primarily through mobile apps and operating systems.
It’s rare to find computer vision use cases tied to a desktop experience. Seeing AI, for example, a service made by Microsoft that uses computer vision to help the blind interpret the world, is only available on iOS smartphones.