Yeah the same basic structure works, but you gotta specify about being open in which topology, with subspaces running around.
For instance take any open covering of F(K) with sets open in Y (this is important because continuity of f is w.r.t. the topology on Y). Pull back, its still an open cover by continuity and set inclusion invariance under pullbacks of continuous maps. K is compact so yeah just pick the finite subcover, then send that stuff back into Y and it will still cover K, but it is finite now, and they are still a subset of the original collection of open sets. (perhaps an argument should be made here about the open sets still being the same you know like why is ).
But yeah it is pretty much exactly the same, just want to stress that you just gotta at least state where the sets are open when dealing with subsets like this generally.