I also appreciate the great discussion that's going on in the comments - you folks post comments faster than I can respond to them, but other commenters are making many of the same responses that I would make, so all's well as far as I'm concerned.
My main point in this post was just that, whatever you think of universals such as numbers and propositions, colors at least are not mind-independent. In TLS, Feser made mind-independence part of his definition of realism:
The view that universals, numbers, and/or propositions exist objectively, apart from the human mind and distinct from any material or physical features of the world, is called realism.... (p.41)
And in his argument for realism, he emphasized that
When you and I entertain any concept - the concept of a dog, say, or of redness.... - we are each entertaining one and the same concept; it is not that you are entertaining your private concept of red and I am entertaining mine, with nothing in common between them.
In his first response, though, he seems to concede both these points. When he distinguishes between "RED" (the sensation or experience) and "red" (the physical properties that produce this sensation), he defines them as follows:
RED: the qualitative character of the color sensations had by a normal observer when he looks at fire engines, “Stop” signs, Superman’s cape, etc.
red: whatever set of physical properties it is in fire engines, “Stop” signs, Superman’s cape, etc. that causes normal observers to have RED sensations...
These definitions both refer to "a normal observer" and so are implicitly mind-dependent definitions. (Unless Feser thinks we can define "normal observer" without reference to the physiological and psychological properties of human beings.) And he states
And even if it turns out that what you call RED and what I call RED are qualitatively different, that wouldn’t make any difference either. For then it will still be the case that every instance of RED in my sense of “RED” is an instance of the same one universal, and every instance of RED in your, different sense of “RED” is an instance of a different universal, and so forth.
But in that case, we wouldn't be entertaining one and the same concept.
Feser calls his position "Scholastic realism,"
...on which universals exist either in the things which instantiate them or in intellects, but where the latter includes the divine intellect, in which they pre-exist as the archetypes according to which God creates.
That they can exist "in intellects" seems to allow for a certain amount of mind-dependence of universals - clearly a more nuanced view than what he presented in the book. I don't understand why this mind-dependence doesn't put him in the conceptualist camp, rather than the realist camp. But it seems that on the basic point - the mind-dependence of colors - we are in agreement.