I currently have some code which captures a still image from the Pi Camera Module and then identifies faces using the haarcascade xml file provided with OpenCV for Python. The code that I am using is the code shown towards the end of this blog post: http://rpihome.blogspot.co.uk/2015/03/face-detection-with-raspberry-pi.html, however it is slightly modified (fully working).
The only problem is that it currently only recognises faces on still images. Is there any way to make it so that I can continuously stream from the Pi Camera directly to OpenCV and then process the faces and display boxes around faces live in a window instead of saving a single frame to a file? I have tried several different tutorials online, but they all seem to not work for me.