refactor(data/set/finite): change type of `set.finite.dependent_image` (#6475)
The old lemma combined a statement similar to `set.finite.image` with
`set.finite.subset`. The new statement is a direct generalization of
`set.finite.image`.
Co-authored-by: sgouezel <sebastien.gouezel@univ-rennes1.fr>