Thanks for your attention to this newbie in CoreImage programming!
I think this must be an old topic since when I met this trouble I
found many similar problems on line. However, there are no perfect
solution for it, and most importantly, they are not feasible to my
situation.
I am now on a video analysis program. We have get the QT movie and
show it in a OpenGLView using Core Video(mainly based on the source
code of civideoDemo). Now we need to do pixel analysis on the frame of
the movie. I am sure that there is no problem on the video->frame
since we have done some composite filter based on the CIImage frame we
get from the movie.
It is a straight forward idea that we convert each frame into a Core
Image object using
[CIImage imageWithCVImageBuffer:currentFrame]
Here currentFrame is defined by:
CVImageBufferRef currentFrame;
and then our idea is to get the pixel information using
NSBitmapImageRep.
One solution of using [[NSBitmapImageRep alloc] initWithCIImage]
failed when the program run into a Exec_Bad_Access error.
I tried another way: draw the CIImage on to a CIContext and then
render it to bitmap data:
