I spotted two errors in the book regarding the residual error equation (Optical
flow, p 470)
The double summation currently has this expression (I(x,y) - J(x+dx+dy))^2, but
I think this should be (I(x,y) - J(x+dx, y+dy))^2 (two parameters are needed
to find a location in J, not one)
Also, in the line below it says that there is a window of size (wx, wy), but
since the summation goes from x=ux-wx to x=ux+wx, the window is actually of
size 2*wx, 2*wy
HTH
Original issue reported on code.google.com by lexvande...@gmail.com on 11 Jul 2013 at 11:29
Original issue reported on code.google.com by
lexvande...@gmail.com
on 11 Jul 2013 at 11:29