Hi, I'm coding an algorithm for background estimation using Intel IPP, and I need to calculate the euclidean difference between pixels (RGB) of two consecutive video frames.
In other words, if Px and Py are the two RGB pixels I need to determine the value:
d(x,y) = sqrt( (Rx-Ry) + (Gx-Gy) + (Bx-By) ).
I see in the manual that there are some functions that can calculate the euclidean distance between an image and a template, but I can't figure out how can I use one of them to fit my needs.
Do you think that using a loop to iterate on each pixel of the two images is much more
expensive in calculation time than using the IPP function?
We do not have such single primitive function. But you may try to implement this algorithm with combination of call for several IPP functions.