Hi all
I've read up on the brief overview OpenCV give for the LBP implementation. However, in practice MB-LBP is used, where pixels are replaced by rectangular regions. These rectangles are used to take the average of pixel regions in order to capture larger structures. However, I've not found any information on how this process is combined with the LBP-Histograms that is stated in the OpenCV documentation. As in order to generate a histogram, you need more than 1 pattern, which is all I can see you yielding from the OpenCV code I've read through.
If anyone can shed some light on how these two processes are combined it'd be much appreciated.
Cheers
Div