How to Make images of Text Appear Larger and Readable for OCR Article
Sometimes there are images of text that, originally, will have very small print in them when scanned or downloaded from the Internet. Examples of such include identification cards that you can scan yourself, print invoices or newspapers that contain labels enclosed in small print, and yes, that fine print you do not have to miss out on when reading a print on-sale coupon. There are also images of PDF documents that people put out on the web, and of course, no screen reading technology can recognize or read. Even when you scan them with your scanner and use adaptive technologies such as OpenBook® to recognize your page, when the print is too small and unreadable for the OCR software, there are steps you can follow to make it recognize better.
1. Probably one of the most obvious causes that the print can be too small can be straightened out with resizing or changing the image’s Zoom or Stretch settings. If your image of text is in a JPEG format, you can work with it very easily. You can just edit the picture file using Microsoft Paint, as described in the following procedure:
A. Open the image containing the text using Paint.
B. if you are using Windows XP, you need to go to the Image menu by pressing ALT+I.
C. Once there, find an option that is called “stretch/Skew…” in this menu.
D. Press enter. A dialog box should appear with four options, as explained below.
E. In the horizontal Stretch option, which is set by default to 100, type a larger value, for example, 500. This will make the horizontal viewing much bigger.
F. In the vertical Stretch option, type the same number, so that all of your edges of the image go at same length and width to be readable by your OCR technology.
Note that this will change the resolution, in pixels, of the image file; therefore, you need to save your image before continuing, usually by pressing CONTROL+S.
Once this is done, and your image is saved with the new resolutions, you can go ahead and try sending it to your OCR software for recognition. The quality of the text plus the handling of junk characters should improve substantially because your image file is now bigger, and therefore, the tiny text that was in the image is bigger.
2. Another option is, if you have a Word document or a PDF file to work with, you cannot easily edit the image using any of the settings described when using Paint. Therefore, simply, if you have a virtual printer driver installed on your system, like the Windows Fax Driver that comes with windows 7 and the Microsoft Office Document Image Writer that comes with office 2003 (can be installed in later versions also), you can send the word or other non-JPEG image-based documents to that printer. All you do is go to your print command on that specific application, and select your virtual printer driver. Virtual drivers usually save their printouts of files with .TIF or .JPJ extensions, so after this point, you can go ahead and apply the steps A through E above under section 1, Working with paint.
If you do not have a virtual printer installed, you can then print out the document on a physical printer, and then scan with a scanner or camera, although for blind people who work or leave alone may have to throw away the page or recycle—which would be the best thing to do in that kind of situation.
Once the page finished scanning, you can then go ahead and apply the steps A through E above when I talk about using paint, that is, if your scanner supports scanning to JPG or other image file formats.
3. Other options: if you do not have a scanner or printer, and think that an image-based document was sent to you (or downloaded) with smaller print, you have two options. If it is an already image file format like JPG or PNG, you can go straight to using paint to edit and stretch the image as described in steps A through E. if it is, say, a word document containing an image with tiny text, you can try and do a print-screen to create a screenshot of the document (assuming it’s in the foreground), and use Paint to paste and edit the image stretching. Then you can use “RoboBraille®”:http://www.robobraille.org as discussed in another article I wrote, and you can see the details on how to use this online OCR service.
Some useful reasons why you want to OCR small print is, of course, fine print you need to look at, but also many newspapers in print have smaller than normal fonts in their pages to make room for colors, pictures, and more stories depending on their regular page size. Also, you may want to read promo-code cards that they gift to you from Amazon or ITunes, which may have small print, and you can enlarge yourself with the options I explained above. Finally, happy print reading!