Improving scan quality for Optical Character Recognition (OCR)

When you use source-document processing features, the OCR process can misread poor quality or otherwise defective scans, resulting in incorrect values within one or more source documents. This topic provides suggestions to help you eliminate variables in the scanning process that can lead to inaccurate data.

To achieve optimal image quality when using source-document processing features, follow these guidelines when setting up your scanner.

  • Use original documents whenever possible. Copying and faxing degrades the quality of characters.
  • Set your scanner resolution to 400 DPI.
  • Set image mode to black and white. If you must use color or grayscale image modes, set your scanner resolution to 300 DPI and use the lowest color depth, 8 bit, instead of 24 bit. color and grayscale images produce file sizes 10 to 20 times larger than black and white images.
  • Set compression to high quality.
  • Contrast and black level should be set to medium.
  • Consider purchasing and installing a scanning application that offers additional settings and functionality to help clean up scanned images.
  • Choose Image > Scan, click the Options button, clear the Reduce image size on disk checkbox, and then click OK.
  • Scan area should be set to 8 1/2 x 11 or letter size (do not use auto adjust or crop settings).
  • Ensure that there are no stains or other blemishes on the documents (such as coffee stains, marker annotations, and so on)

Note: If, during tax season, you receive a supported brokerage statement that hasn't had its data extracted, retransmit the document on the following Monday. FileCabinet CS implements OCR updates every weekend to correct recognition problems caused by format changes.

Was this article helpful?

Thank you for the feedback!