Hi Marco,
The function “imfuse” by default use “falsecolor” algorithm, so the green and magenta color you able to see are the regions where the intensities are different of the image I1 and I2 provided to the function. The first and the third channel of the output “Fused” has the information about the magenta color, while the second channel gives the information about the green color.
One easy way I can think of to address your concern is to extract magenta color channel from the output image and use it. I understand you need to work something more on top of it. But atleast extracting that channel gives an head start or some idea about the region of pixels which has color magenta.
I hope it helps you.
