One reason why understanding vision is so difficult is that we who are attempting to understand the process are so deeply embedded in the phenomenology of perception: We know what it feels like to see. We look out and see the world, and we cannot escape the impression that what we have in our heads is a detailed, stable, extended, and veridical display that corresponds to the scene before us. Of course, most of us have also seen enough examples of so-called "optical illusions," so we are prepared to admit that what we see is not always what is truly the case. Yet at the same time we have much more difficulty shedding the view that in our heads is a display that our inner first-person self, or our cognitive homunculus, observes. There are other phenomena relating to our experience of seeing a "picture in our head" that are even more problematic. These include the similar experience that we have without any visual input: the experience that accompanies mental imagery or visual thinking. The more we analyze what must be going on and the more we examine the empirical evidence, the more puzzling the process becomes and the less tenable our intuitions. Indeed, we find not only that we must dispense with the "picture in the head," but that we must also revise our ideas concerning the nature of the mechanisms involved in vision and concerning the nature of the internal informational states corresponding to percepts or images. What can never serve as a theory of vision is a theory that says that vision creates a copy of the world inside the head, as the Kliban cartoon in figure 1.1 suggests is the case with a cat. The understanding that this sort of theory will not do is what makes this cartoon funny. Yet it is nontrivial to say what exactly is wrong with a theory that even remotely resembles this sort of story (Pylyshyn, 2003, pp. 2-3).Now, Pylyshyn (like everyone right?) rejects the "homunculus" theory of perception, just as does Gibson. But, Pylyshyn and Gibson reject it for very different reasons (which I won't go in to.) The point is that (Gibson's contentions notwithstanding) one can start with an image projected on the retina, but still not end up by postulating a homunculus.
Note as well how Pylyshyn, like Noë, rejects part of what is meant by an imagistic theory of perception, albeit again for very different reasons.
Pylyshyn, Z. (2003). Seeing and Visualizing: It's not What you Think. Cambridge, MA: MIT Press.