What method would be best suited to gather data from images where the name of the data label might change from image to image but still able to be defined as a list?

An accounts payable department needs to collect and gather information from the receipts of its employees as part of their expense reporting process. They have chosen IBM Datacap as the tool they want to use to automate large portions of this process.

What method would be best suited to gather data from images where the name of the data label might change from image to image but still able to be defined as a list?
A . Zonal
B . Click N Key
C . Fingerprinting
D . locating by keyword

View Answer

Answer: D

Explanation:

Keywords are labels that accompany the data you want to find on the form.

Locating by keyword ranks third behind the zonal and regular expression methods in terms of reliability because you have to know what the keyword is and recognize the keyword correctly. You also have to anticipate the location of the actual data, relative to the position of the keyword, for extraction.

Keyword extraction is made easier by using keyword lists, which are text files that contain many different keywords that might be used as labels around the data you want to extract. For instance, a keyword list could contain Total, Total Due, Pay this amount, and similar words or phrases that occur in close proximity to the actual data.

Note: There are four ways of extracting data from an image. They are ranked in order of preferable practice: Zonal, Locate with regular expressions, Locating by keyword, Click N Key

References: IBM Redbooks, Implementing Document Imaging and Capture Solutions with IBM Datacap, Second Edition (October 2015), page 125