Hierzu sollte zunächst, the webcam connected. The webcam should have a resolution of 320x240 pixels or higher resolution. connected
After the WebCam is, the program can be started HandVU_OpenCV.
After starting the software that displays the camera image and a rectangle in which the hand must be positioned until the hand is detected.
One should be the back of the hand, right hand, place it on the rectangle, since the hand position is best detected.
be seen in the other pictures, as should be the right hand position, the hand gesture recognition zu gewährleisten:
Man sollte auf gute Lichtverhältnisse im Raum achten.
Der Raum darf weder zu dunkel, noch zu hell sein. Durch Sonnenschein, wird beispielsweise, die Erkennung der Hand durch die WebCam beeinträchtigt.
Auch Hautfarbeähnliche Hintergründe sollte man vermeiden.
In meinen Tests habe ich festgestellt, dass beispielsweise helle Holztüren als Hintergrund, nicht für den Einsatz von HandVU geeignet sind.
Bei dunklen Hintergründen the hand was detected very quickly.
The hand is not detected until the large rectangle will disappear and be distributed to the recognized hand, white dots.
This is the big white dot, the center of the computational rectangle.
will now move at the back of the hand, the hand with the white dots followed.
Now they have turn the possibility of the hand by very fast to point the palms towards the camera.
It also has the ability to hand exchange with other objects. For example, I tried it with a cup of glass (and see picture). The object tracking of the cup has works very well.
ie not to say that the object can be used at start of HandVU. There must continue to be hand held in front of the webcam, until the palm is recognized.
The hand is also detected very well if people are moving in the background.
images of the test results:
Website:
http://www . movesinstitute.org / ~ Kolsch / HandVu / doc / index.html