PsychKinect – Control and access the Microsoft Kinect depth camera.

This is a high level driver to allow convenient access to the Microsoft
Kinect box. The Kinect is a depth-sensing “3D camera”. The Kinect
consists of a standard color camera (like any standard USB webcam)
to capture a scene at 640x480 pixels resolution in RGB8 color with up to
30 frames per second. In addition it has a depth sensor that measures the
distance of each “pixel” from the camera. The Kinect delivers a color
image with depth information which can be used to infer the 3D structure
of the observed visual scene and to perform a 3D reconstruction of the

See KinectDemo and Kinect3DDemo for two demos utilizing PsychKinect to
demonstrate this basic functions of the Kinect.

PsychKinect internally uses the PsychKinectCore MEX file which actually
interfaces with the Kinect.

The driver is currently supported on Microsoft Windows under
Matlab version 7.4 (R2007a) and later. It is also supported on
GNU/Linux with Matlab or Octave and on Intel based Macintosh computers
under OS/X with Matlab or Octave. The driver supports all versions of
the Microsoft Kinect on Linux and OSX, but currently only the original
XBOX-360 Kinect under Microsoft Windows.

To use this driver you need:

  1. A Microsoft Kinect (price tag about 150$ at December 2010).
  2. A interface cable to connect the Kinect to a standard USB port of a
    computer - sold separately or included in standalone Kinects.
  3. The free and open-source libfreenect + libusb libraries and drivers
    from the OpenKinect project (Homepage: )

You need to install these libraries separately, otherwise our driver will
abort with an “Invalid MEX file” error.

Type “help InstallKinect” for installation instructions and licensing


Most functions are part of the PsychKinectCore() mex file, so you can get
help for them by typing PsychKinectCore FUNCTIONNAME ? as usual, with
FUNCTIONNAME being the name of the function you want to get help for.


PsychKinect(‘ApplyCalibrationFile’, kinect, calibFileName);

  • Load Kinect calibration from a .yml calibration file ‘calibFileName’, as
    created by rgbDemo software and apply it to Kinect with handle ‘kinect’.

kobject = PsychKinect(‘CreateObject’, window, kinect [, oldkobject]);

  • Create a new kobject for the specified ‘window’, using the Kinect box
    specified by the given ‘kinect’ handle. Recycle ‘oldkobject’ so save
    memory and resources if ‘oldkobject’ is provided. Otherwise create a new

Do not use within creen(‘BeginOpenGL’, window); and Screen(‘EndOpenGL’,
window); calls, as 2D mode is needed.

kobject encodes the 3D geometry of the scene as sensed by the Kinect. It
corresponds to the currently selected “3d video frame” from the kinect,
as selected by PsychKinect(‘GrabFrame’, kinect);
kobject can then by accessed directly (it is a struct variable) or passed
to other PsychKinect functions for display and processing.

PsychKinect(‘DeleteObject’, window, kobject);

  • Delete given ‘kobject’ for given ‘window’ once you no longer need it.
    During a work-loop you could also pass ‘kobject’ to the next
    PsychKinect(‘CreateObject’, …); call as ‘oldkobject’ to recycle it for
    reasons of computational efficiency.

Do not use within creen(‘BeginOpenGL’, window); and Screen(‘EndOpenGL’,
window); calls, as 2D mode is needed.

PsychKinect(‘DrawObject’, window, kobject [, drawtype=0]);
Draw the 3D scene stored in ‘kobject’ into the ‘window’ selected. Use the
current OpenGL settings for this. ‘drawtype’ is optional and defines kind
of rendering: 0 (the default) draws a colored point-cloud of all sensed
3D points. 1 draws a dense textured 3D surface mesh.

This function must be enclosed between Screen(‘BeginOpenGL’, window);
and Screen(‘EndOpenGL’, window); calls, as 3D mode is needed.

Path   Retrieve current version from GitHub | View changelog