To modify the extended settings for the Brainware Layout Classification engine, complete the following steps.
- On the Extended tab, to specify the number of horizontal zones each document is divided into, modify the Horizontal covering frequency value.
- To specify the number of vertical zones each document is divided into, modify the Vertical covering frequency value.
-
Complete one of the following substeps.
- To treat “123” and “987” as the same sequence of digits of the same length, select Do not differentiate between digits. Use this option when most numbers on your documents do not have a specific meaning, for example, in the tables of a typical invoice document.
- If the documents contain specific sequences of digits that can be critical for layout classification, such as bank account information specific to each vendor class, clear Do not differentiate between digits.
-
Optional. Select
Attach delimiters to words.
Note: In most cases, delimiters such as “,”, “.”, “/”, “;”,are essential to distinguish between words. A US date format, for example, may be better treated as a single differentiating word – “00/00/000” - in the classification’s surrounding compared with 3 “noise” words “00”, “00” and “0000”.
- Save and relearn the project to apply the new settings.