locked
Doubt with HD Face Basics RRS feed

  • Question

  • Hi,

    How is possible  coordinating the MeshGeometry3D of the example HD Face Basics with the real position of my face?

    There is an offset between color video (where is located my face) and the mesh (contained in a viewport3d) that I'm trying to put just on my face.

    I've tried to do this with the coordinate mapper but I couldn't get run it. I don't know how to do this.

    I'm using Visual Studio and C#

    Sorry for my bad explanation and thank you. 

    Thursday, May 21, 2015 4:18 PM

Answers

  • HD Face Basics points exist in camera space, while color space points exist in color space.  The coordinate mapper function you are looking to use is MapCameraPointsToColorSpace (or any of the related camera to color space converter methods): https://msdn.microsoft.com/en-us/library/windowspreview.kinect.coordinatemapper.mapcamerapointtocolorspace.aspx

    If you're trying to map the HD Face Basics renderer onto a color camera image, you may have to choose a different rendering strategy. MapCameraPointsToColorSpace will return ColorSpacePoints, which are X and Y coordinates based on a 1920x1080 pixel image. The HD Face Basics sample is rendering the face based on X, Y, and Z points of a GeometryModel3D.  You'll need to do multiple conversions or clever translations to get the FaceModel into color space.
    Friday, May 22, 2015 10:54 PM