I'm developing some methods for processing images. I have applied seven methods (M1 to M2) in 10000 images and measured the effects of each method in each image according to two performance measures. After that, I created two heat maps representing the pearson correlation between each pair of methods (considering their effects on the 10000 images) for the two measures. My goal was to detect some significant differences in the behavior of the methods. It is important to notice that I used pearson correlation because the range of the magnitude of the effects of each method in the images is different. Thus, if two methods have a perfect correlation, but with different magnitudes in their effects, they are not distinguishable (they are producing the same qualitative result).
However, I'm not sure about how to proceed. How can I determine that the differences between the correlations are significant?
I'm including the heat maps of the pearson correlation for the two performance measures.
