I was reading on wikipedia about the Lucy-Richardson algorithm and its equivalent iterative function:
$$ u^{(t+1)}=u^{(t)}\cdot \Big(\frac{d}{u^{(t)}\otimes p}\otimes \hat{p}\Big) $$
where d is the observed pixel, $u^{(t)}$ is the current iteration, $u^{(t+1)}$ is the successive iteration, p is the point spread function, and $\hat{p}$ is the flipped point spread function.
I understand, at least in an algorithmic sense, what everything is except the the flipped point spread function. Wikipedia explained it as this:
$$ \hat{p}=p_{(i-n)(j-m)}, 0\leq n, m\leq i, j $$
Again, all well and good, assuming one knows what m, n, i, and j are. In this context, i and j are the xy-coordinate of the pixel u. However, my trouble lies with m and n: what range are they on? Is any value fine as long as it's greater than 0, or less than i, respectively; or is there some relation to j or i?
Anyway, if someone could explain in general what flipping the point spread function does, and what these new variables actually mean in relation to the old ones, I'd really appreciate it.
