Migration of internal images fails due to `common` plugin not labeling deployments

Due to missing annotations/labels, the restoration of imagestreams is failing to run through the proper image restoration process.

This is because of a regression introduced to how the common plugin gets applied to a subset of resources: https://github.com/openshift/openshift-velero-plugin/commit/5f2bd726927f7418c1f9df6c2e962cec40c689dc
We think the fix is to apply the plugin to `deployments.apps`.