OCR Documents with a Camera
What does OCR stand for? OCR is Optical Character Recognition. OCR is the conversion of images of text, e.g. a scanned document, into plain text, like you get from typing in a word processor. Document imaging is often enhanced by optical character recognition, allowing documents to be indexed and searched.
|
Before PhotoNote
|
After PhotoNote
|
PhotoNote was used to prepare the above image for Adobe Acrobat 8 Standard. Acrobat was used to create a searchable PDF. Download the sample pdf file.
You need a program to perform OCR on images. You can try the built in OCR of PhotoNote or use PhotoNote to prepare images to get the most out of 3rd party software.
OCR software is sensitive to the qualities of an image. A poor image will produce many OCR errors, if the text can be read at all. Use the tips below to take good photos and Here we'll discuss the major issues and ways to improve your images for OCR quality.
Lighting
One of the key elements for all of photography is lighting. This is especially true when you with to OCR your images. If using a flash be sure to avoid glare on the page. To solve this problem take a picture from an angle and use PhotoNote to correct keystoning. If not using a flash best results will be had with a tripod to hold the camera steady.
Image is Blurry
Reduce blur by using a tripod. Alternately, brace the camera on a table, a stack of books, or whatever is around. Using a flash will help, but make sure it doesn't produce a glare. Finally, image sharpening may help if unable to reduce blur in the original images.
Perspective
The image of the document will be skewed (called keystoning), when the image is taken from an angle. Perspective correction will make the document look flat, like it was scanned.
OCR Engine
PhotoNote uses Tesseract OCR. Google supports current development on Tesseract, we have high hopes for ongoing improvements in OCR performance. Tesseract currently outperforms free ocr programs.
