Scanning text for best OCR results
When you scan in a text document, use these recommended settings for the best results:
|
Item type |
Resolution |
|
Documents |
300 dpi black-and-white |
|
Documents of poor quality |
400 dpi black-and-white |
You can use PaperPort’s built-in OCR software to copy text from scanned text documents so that you can use and edit the text in any text editing, word processing, or spreadsheet program.
This process is often referred to as a conversion process, because the OCR software reads, translates, and converts scanned characters into actual text.
If PaperPort exists on a computer with OmniPage, its OCR services become available and amplify the power of PaperPort. See Using OmniPage with PaperPort. For a detailed list of OCR features in OmniPage, see www.nuance.com.
Notes
The OCR process requires textual images that are scanned at 150 to 600 dpi.
When translating scanned text, the OCR process is not always perfect. For most documents, depending on the quality of your original, you can expect up to 98 percent accuracy with PaperPort OCR software. Some documents, however, are hard to OCR. If your document has fine print, you might find that scanning it at 400 dpi produces better results. Be aware that scanning at a high resolution, such as 600 dpi, will generally not improve OCR results and may actually degrade the OCR quality.
OCR does best when the file is in a lossless format, such as PaperPort Image (.max) files set with Best Quality, Windows Bitmap (.bmp) files, or TIFF files. OCR results with JPEG files and with PaperPort Image files set with a lower quality are dependent upon the amount of compression.
Tips
For best OCR results, start with a clean original.
If an image item contains handwritten notes or lines, the OCR software tries to analyze this information and might make an incorrect conversion. For this reason, you should always check the converted text after the OCR process is finished.
Maps, handwriting, pictures, logos, and similar items do not produce good results when run through the OCR process.
If a scanned text image also contains lines, artwork, or handwriting, display the image in the Image View window, cut the extra lines or shapes, and then use the OCR process to extract text from the image.
//