You can use Automator.
Open Automator.
Choose Application from the templates.
Add a Ask for Text action. Tick 'Require an Answer'
Add a Create Banner Image from Text action.
Tick 'Use text input as image filename'.
Open the actions options options.
Tick 'Show this action when workflow runs'
Add a Lopp Action.
Control + click on the Loop Actions Title bar and select 'ignore input'
From the Loop Actions's drop down menus, select 'Ask to Continue' and Use the 'current results as input'
Save and run.
You will then be prompted to enter your text.
Then after you ok that you will be prompted to Choose you font.
Use the font size to determine the image size.
example 64 in Helvetica Bold gives me about 46x77
The same font but at size 300 gives me 177x360.
In each case I am only using 1 character.
If you want to do it from 'Selection'
Open Automator.
Choose Services from the templates.
Set the 'Sevices Receives input' to :rich text * and
in* any application
Add a Create Banner Image from Text action.
Tick 'Use text input as image filename'.
Set the Text Style drop down menu to : 'Use Style of Rich Text input'.
Optionally you can add a 'Reveal in Finder' Action
The Size is set by what ever the original fonts are.
To use the Service.
You can select some text, and control click to get a contextual menu. Listed is the services sub menu. Under this you should find your new service that will run on the selected text.
Best Answer
Skitch
Skitch is a very useful app for making annotations on images, and whilst by default it saves the files using the format that they are imported as, it can save it as a Skitch PNG which keeps the ability to modify the annotations, or the files can be kept inside its own library.