MO vs. VB for ketone dyes

I am thinking more about photophysical properties of ketocyanine dyes. I have puzzled through the basics of the molecular orbitals. Some material on this site is useful, including a visualisation of the molecular orbitals of formaldehyde. A key point is that the HOMO is a non-bonding orbital centred on the O atom and lying perpendicual to the C=O bond.



Thinking in the alternative resonating valence bond picture there will be three alternative Lewis structures
O
||
C -R
|
L

O-
|
C -R
|
L+

O-
|
C -R+
|
L

The extent to which the lower two structures contribute will increase the C-O bond length and reduce the C-O stretch frequency.
It should be possible to describe the low lying excited states in terms of a complete active space with 4 electrons in 3 orbitals (a pi* orbital on the C=O bridge, and a pi orbital on the left and right fragments).