In the Image Target example in the Android sample app (specifically in ImageTargetRenderer.java), I can see that Vuforia has kindly handled the model view projection matrix for us, after the image target has been detected and tracked. I know the model view projection matrix is used in OpenGL for rendering, i.e. projecting 3D coordinates in world space up to clip space and screen space.
What I want to do is, to rectify the distorted image target captured on the phone screen, after it is detected, and further analyze the rectified image (probably using OpenCV). As far as I know, knowing the homography between the distorted image and the rectified image should do the trick. So the question is, how can I relate the model view projection matrix, given by Vuforia, and the homography between the two planar image?
Thanks a lot.