At atomic scale sometimes there are perfect cirlces. The argumentum a silentio fallacy is bidirectional in this case.
Now go with your atomic imperfections and external influences and take a nap, because it must be exhausting to be so repellent.
The way I see it, the circle is an abstract idea which resides within the language that we use to share our understanding about the patterns we observe in the external world. Those patterns may conform to the abstraction to varying degrees but they are fundamentally not the same thing. Mathematics is a very precise and formal way to communicate this type of knowledge, but that doesn't mean the ideas therein exist objectively. This contrasts with the views of someone like Max Tegmark, who has the impression that the universe is inherently mathematical.