Enriches semantic Text nodes with colors measured from one rendered view capture.
Returns unchanged nodes when the rendered colors cannot be measured confidently.
Enriches nodes using a rendered bitmap whose origin matches the semantics host.