unusual convention for the image center position
After every call to a cor-finder method, the result, which is an absolute position, is transformed to an offset, using this formula res = res - scan.dim_1 // 2 thus obtaining the offset from a supposed center. This gives an offset which, for even sizes, is half a pixel smaller that the usual convention. I suppose that this cancels out when the absolute center is written again, if the so calculated offset is added to the same unusual convention for the image center. But the all may be misleading in the manual search interface where the input is the offset from a (variable convention) center.
Edited by myron