[enhancement]: Consistent Character Generation From Text

### Is there an existing issue for this?

- [X] I have searched the existing issues

### Contact Details

_No response_

### What should this feature add?

This feature would add the ability to generate and name a character (or object) for later re-use.

A recent research paper suggests a method for generating consistent characters solely from a text input (without any intermediate user intervention). This would be hugely useful for character work if it were implemented in InvokeAI.

https://arxiv.org/abs/2311.10093
https://omriavrahami.com/the-chosen-one/

### Alternatives

Manual processes exist, but involve generating and winnowing images and then training a model on them for each character. The research proposes a much more straightforward, text-only, single-input approach.

### Additional Content

There is a larger prompt _composition_ question that this raises: one could imagine a UI where a character was generated and then could be inserted into scenes. In this case, the prompt for the scene and the prompt for the subject(s), and object(s) would each be encapsulated and (re)composable.

In such a case, a workflow would involve first creating the various components (setting, characters, objects) individually and then combining them. (This is only possible if these components are stable, which this research suggests is possible.)

In this case, it would be useful not only to be able to generate consistent characters/objects but also to name them and re-use them later.