I have a great many postcards scanned, and would benefit greatly from being able to OCR the text not displayed perfectly horizontal. I've read the other thread on this, but it's incomplete. http://forums.adobe.com/thread/988537
Does anyone know a good way to do this? 
Scan all the postcards.
Use a common orientation.
Example - scan all to obtain an output image with the landscape orientation of the image upload you provided with your post.
The orientation can be landscape or portrait. Just ensure the initial orientation of all are the same.
Use TIFF as the scanner output image and then use Acrobat to create PDFs from these.
Or configure Acrobat to work with the scanner software to end up with PDF(s).
You can processe the individual PDFs via Actions or you combine the many PDFs into one PDF and process this single, many page PDF.
OCR the PDF(s) with all in the same orientation.
Save the OCR results with the file(s).
Rotate the PDF(s) 90 degrees and OCR again.
Save.
Rotated to final desired orientation.
Save as.
Build Catalog index / indexes.
Build the catalog mount PDF and via Document Properties > Advanced connect PDF to PDX to auto-mount Catalog index upon opening of PDF. Once PDF is open engage in Advanced Search of collected, OCR's postcards.
Notes:
--| Catalog index only works on the local computer or network space.
--| OCR is for "typed" text (fonts) -- *Not*handwritten stuff of any flavor (for that see ICR).
Be well...