Optical Character Recognition (OCR)
This sample demonstrated the capabilities of a typical OCR application with
varying conditions. Notice that the accuracy of the document text remains at a
100% unless there are some interference like background images or deteriorated
text. The basic rule of thumb is that if the edges of the fonts / text are
clearly visible, the OCR will have a higher degree of accuracy. The following
conditions would result in the accuracy dropping significantly:
- Irregular Surface Conditions
- Font / Text Edges are not clear or defined
- Background noise or images
- Paper quality
- Image quality
The following link presents you two files. One without the OCR (the orginal scanned image from a test sample) and the other with the OCR. There is also a word file which can be referenced in order to inspect the aacuracy.
PDF without OCR - Original scan from sample document
PDF with OCR - Text searchable document
Text File (MS Word) - Uncorrected reference text document exported from the OCRed PDF
Notes on Test Conditions
Standard A4 recycled paper was used since the purpose of this demonstration is to exhibit the capabilities of OCRing. The scan was conducted at a resolution of 300DPI with image settings of Black and White / Text. You will be needing the latest version of Adobe Reader to view the above PDF documents.