The MatchingMapMaker (M3) is a MATLAB function used to detect shifts between two images using convolutions. The function will operate a weighted sum of absolute errors for each lag vector given as input.
The main code is written in C/C++, but the code is interfaced to MATLAB using mex.
The function mmm use 4 parameters:
- an odd size kernel, which can be full of one to compute the Sum of absolute error, or a normalized version to have the mean
- reference image
- compared image
- a matrix, where each line represents a potential lag vector.
and return for each pixel:
- the lag with the lowest error
- the lowest error
[lagIndex,quality]=mmm(kernel,ref,shiftedImage,lagVector,'Algo');
- test with
mex -setup c++
that the mingw-w64 compiler is installed, in other cases install it from here - run
compile
- run
test
(to test if it works! )