I ask this question because I am having success using vImage in the
The very large arrays of which I speak are merely image data in
memory. I need to take one image and subtract it from another.
(imageDataA[i] - imageDataB[i] = imageDataC[c])
right now I just do it via indirection (pointers) in a for loop. The
images are in the range of 8 megapixels so the calculations are in the
millions per image (time many images)
Unfortunately this simple subtraction problem is as complicated as it
gets for this one part of the app.
The data comes to me from a camera as an imageData array. (unsigned
char) To use vDSP (as I understand) I would have to "vectorize" the
imageData array and then perform my simple math operation and then
translate it back into unsigned char.
For floating point images, you can probably use vDSP_vsub(). If the
rowBytes parameter is the same for both images (and is divisible by
four) you can just use vDSP_vsub for the whole image. If the rowBytes
are different, you'll need to call it once per scanline.
Do not post admin requests to the list. They will be ignored.
PerfOptimization-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden