The molecular orbitals were calculated from a minimal basis with AM1 parametrisation in Arguslab. They are pretty close to the ones of nitrogen. The difference is that lower energy MOs are rather O-centered and higher energy MOs rather C-centered. This makes sense because of the fact that the atomic orbitals of oxygen are lower in energy. It can be seen that the HOMO and the LUMO  are centered on the carbon. This causes the strong σ-donor and π-acceptor characteristics of CO.
The table has the LCAO coefficients for the different orbitals. The color coding shows if orbitals have a bonding or antibonding interaction . It can be seen the the lowest energy AO (O-2s) has the strongest impact on the lowest energy MO, and that the highest energy AO (C-2p) has the strongest impact on the highest energy MO.
CO is usually drawn with a formal positive charge on the oxygen and a negative charge on the carbon.
Formal charge and polar bonds work in opposite ways. This results in a very small dipole moment of .1 Debye. Its direction was argued over for a while. Now theoristical chemists are pretty sure that at equilibrium distance the carbon atom is the negative end.
 Theorists always seem to be kind of careful when talking about unoccupied MOs. In order to sound serious I have to do that, too: One has to remember the fact that unoccupied MOs are even less real than occupied ones. For some reason they are pretty good if you use a minimal base. But as soon as you move from plain LCAO to using more basis functions you will just get a lot of "virtual" MOs close together in energy with no physical significance.
 For σ-MOs the color corresponds to the sign that the wave function has in between the nuclei, for π-MOs it corresponds to the sign in the positive x-direction of the atom.