Why the autocorrelation matrix is an estimate of the inverse of the covariance matrix of the corner position?

[javierttgg]

Hi @pengsongyou , amazing work :) Thanks for open-sourcing it!

If I may, I have a question regarding the following paragraph of the paper:

Consider a corner point extracted in an image; the uncertainty of its position can be estimated by computing the autocorrelation matrix C for a window of a given size around the point (see for instance [5]).
Concretely, C is an estimate of the inverse of the covariance matrix of the corner position.

I am struggling to understand this. Given that the autocorrelation matrix can be computed as explained e.g. in this tutorial). I don't see where this result of being the inverse covariance matrix comes from.

I would highly appreciate if you could shed some light on this,
Thanks in advance!