My problem is as follows:
I have a 2D image (u,v), with each pixel representing the value of the 2D cross correlation function (ccf) at that place. I am trying to determine to what degree I can be confident that the ccf peak is in that region. Specifically I have identified the peak pixel (I, J), around which all other pixels have lower values, like a 2D gaussian function. I am drawing a circle of radius R. My question is, how long must R be, so that I achieve a 95% confidence level.
I would suppose that the answer is something along the lines discussed here, but this is for 1D.
Any help/answer is greatly appreciated.
