[cairo] A better image comparison tool

[Jeff Muizelaar]

On Mon, Sep 22, 2008 at 02:53:57PM -0400, Antoine Azar wrote:
> I am considering a couple approaches. The first one I want to try out is a
> pretty simple square sum of differences or sum of absolute differences.
> There are more fancy algorithms like cross correlation and mutual
> information,  which I'll probably try out too. We can then set custom
> thresholds for what should appear as "OK", "Warning" and "Error". 

Mutual information is probably not going to be a useful metric. It's
most useful for situations where you want to compare images that have
different luminance scales like multi-modal image registration. For
example, an inverted image would have a large amount of mutual
information with the reference image but we probably do not want it to
pass.

I also don't think mutual information would work well with very small
images. In fact, cross correlation might have a similar problem.

-Jeff