Basically, I would like to "cut out" a trapezoid portion of an original image (having the same image with only the trapezoid portion remaining and the rest of it being filled with black would also work for me), and then have it shrunk so that the trapezoid portion fits into a much smaller rectangle image (in essence, just have it resized so that the wider trapezoid part of the image is shrunk more than narrower one) and have that applied to a real-time camera feed...
Since I'm a newbie to image processing on Windows Phone, but having read and heard the Nokia Imaging SDK is greatly optimized, I thought I'd use that instead of trying to access the image and its particular pixels (if that's even possible with the plain WP8 SDK) instead...that's in case the described amount of processing could be done on a regular Windows Phone device with at least 10fps..could you give your opinions? How to best do this? I'm thinking of some rotation filters applied and then cropping the results successively (XY degrees to the left, crop rectangle of orifginal size, rotate back to right by 2*XY degrees and crop again, and then just rotate back to right by XY degrees to get the trapeziod - if the rotation and cropping fills out the missing parts with empty fills).
I'm attaching the illustration I made of what I'd like to do if it helps better explain what I'd like to do.
Hope it helps