Announced Windows Vision Skills (Preview)

Original author: Varsha Parthasarathy
  • Transfer
Recently, a preview version of Windows Vision Skills was announced , which is part of the NuGet suite of packages that makes it easier for application developers to solve a whole range of computer vision problems with a simple API set.

image
Figure 1 - From left to right you see in action: Object Detector, Skeletal Detector and Emotion Recognizer.

With Windows Vision Skills:

  • An application developer can use the out-of-the-box WinRT APIs to add off-the-shelf vision skills, such as object detection, skeletal detection, and face sentiment analysis, to their Windows applications (.NET, Win32, and UWP).
  • A computer vision developer can use hardware accelerated frameworks on Windows devices, packaging his solution as a Vision Skill and not worrying about low-level APIs.

image

All Windows Vision Skills inherit base classes and interfaces from Microsoft.AI.Skills.SkillInterfacePreview . This open framework can be easily expanded to work with existing machine learning frameworks and libraries such as OpenCV.

Windows Vision Skills complements existing Windows support for outputting ONNX models using WinML for local output. The framework allows you to create intelligent applications using the platform optimization capabilities.

Get started now!


Play with the samples provided or learn how to create your own Windows Vision Skills projects using our tutorials .

Do you have an idea for a new feature or would you like to send feedback? Find us on github .

Also popular now: