In the previous article, we used EMGU to identify people's faces, when the Kinect SDK version was 1.0, and after the release of the 1.5 version of the SDK in May, we were able to use the Kinect directly to implement face recognition without the need for Third-party class libraries.
SDK1.5 new Face Recognition class Library: Microsoft.Kinect.Toolkit.FaceTracking makes it easy to face recognition in Kinect, the source code of the class library is also in developer Toolkit. You can also open the run or view the source code in the Developer toolkit with an example of face recognition.
This article uses a simple example to demonstrate how to use the face tracking function in Kinect. The simple feature of this example is to use WPF to draw a simple human face map, and then use facial tracking to move the diagram, which uses only some simple information in the Kinect to get the face, to demonstrate how to use the Kinect for facial tracking, so if you want to achieve a three-dimensional display of the face , or draw a face triangle surface, you can refer to the Developer toolkit in the example.
I. Preparation BEFORE development
To use facial tracking, the Kinect SDK version should be at least 1.5, the latest version is 1.6, you can refer to the previous article download installation, the Kinect SDK and the Kinect Developer Toolkit to install together, My machine is loaded with the latest SDK1.6 version.
Installing the Kinect Developer Toolkit installs Kinect Studio, some c#/vb.net/c++ application samples, source code, and two class libraries for facial tracking FaceTrackData.dll. FaceTrackLib.dll includes 32-bit and 64-bit versions, which should all be under the directory when they are installed: