Hello, Training the VJ algorithm to detect faces

i have a final project for image processing, but i am stuck. i used the viola-jones algorithm but i don't know how to train it to detect either faces or heads from the screenshots. ANY HELP PLEASE !!

Answer 1

[Walter Roberson]

I would not train it to detect "either", I would train it independently for the two and then combine the results. If you only want it to find faces on heads then extract the detected heads and look for faces on that. If you want to find faces outside of heads (e.g., a closeup where the edge of the head might not be visible, or you might want backs or sides of heads) then run both classification steps and you can prune overlaps if you want (but watch out for one object simply being in front of the other.)

[Samah Hussain]

YES i would like to detect faces alone not the two at the same time.. is there any code can help me or some steps to do ? Thank you

[Walter Roberson]

Example 1 at http://www.mathworks.com/help/vision/ref/vision.cascadeobjectdetector-class.html