Yes, you can. It's a combination of settings in how the documents get pulled from the repository and how they get identified in Quick Fields.

Set your Quick Fields session to use Laserfiche Capture Engine to pull documents from the repository. Under "Document Content" in Capture Engine's properties, check "keep each entry as a separate document"

In the document class, don't set any first page identification criteria and leave "last page identification" set to "when the first page identification conditions are satisfied". This way, documents will keep the structure they already have in the repository. Set the name of the document to the %(Name) token to keep their names from the repository and set their location to %(Path) so they'd be sent back where they came from. Under Document Merging in the document class options, you can choose what happens when they're sent back and how Quick Fields should handle field data. You can choose to send just the images back and discard any metadata changes coming from Quick Fields.

Then add any processing you want to do to documents to the Page Processing section of your Quick Fields document class.