Strictly as a Windows user, I don't understand those commands but I can do something similar using Nitropdf, FastView Image Resizer and Acrobat. I have tried it once, I extracted all images using NitroPDF viewer, Resized them using FastView Image Resizer and reassembled using Acrobat. I managed to reduce the file size significantly. The only issue was, if I decide to OCR again, it won't be as good (Its not always needed for my work but sometimes its handy).
The reason, I was asking for a programs that does it 'within itself' is because if I understand correctly when you OCR in Acrobat as searchable image, it 'puts' a layer of text on the image. So I was just curious if it will be possible to keep that 'text layer' as it is and resize the image underneath. Please correct me if I am wrong. There are a couple of programs I found while searching but not too sure how they work.
Thanks for taking time to explain the things (twice). The only problem here is, I don't have the original scans to go through the procedure you recommended but I have made a note of it for future scans.
In some cases, my pure goal is to have electronic version of the document/book for backup or quick printing or reading on ipad on the go purpose. I do not need high resolution pdfs in that case. Since I wasn't well aware of scanning issues in the past, I scanned all at highest possible resolution I could afford and work from there. Now after about 2000 books in Calibre, I realize that its wasting space.
I wonder, if I decide to go by the route Slex suggested, what format is best jpeg or png (and no plans of OCR)? I have heard arguments on both the sides in the context of medical imaging (but I didn't understand a word), just felt like those Nikon and Olympus folks were trying to tell me they are smarter than me