I was wondering if a port of OpenCV to Google Native Client would be feasible? Additionally if it were would I have to go through JavaScript/HTML5 to gain the ability to use a webcam, or does OpenCV implement its own code to access webcams?
P.s I just want to track fingers a heads, is OpenCV overkill?
OpenCV uses platform specific API for camera access. So you need to implement specific VideoCapture back end for it.
How much of an overhaul would you say that is for some who has no experience with the library? And do you think it would be easier to implement the algorithms I described myself, rather than port the library?
You can build and use OpenCV without highgui module. It simplifies you task and allows you to use algorithms from OpenCV.
Update: to disable Highgui module add -DBUILD_opencv_highgui=OFF to cmake command line. You can disable any module using this approach.
Asked: 2013-02-07 18:22:15 -0600
Seen: 318 times
Last updated: Feb 08 '13
