Xerox DocuMate 4799, Руководство пользователя

Страница: 102 / 203

Scanning and Configuring One Touch
Xerox
®
DocuMate
®
4799
User’s Guide
6-94
– Use languages and dictionaries to improve accuracy —select this option to automatically check the
validity of the recognized words. An OCR engine looks at each letter or symbol on the page individually,
then “guesses” what the letter or symbol is based on the shape. Therefore, the OCR engine may have
more than one guess for a particular letter or symbol, a “best guess”, a “second-best guess”, and so on.
This option tells the OCR engine to look through the dictionary to validate its best guess for the letters in
that word; if its best guess is not in the dictionary, it checks for the second-best guess, and so on.
For example, if the word “house” appears in the original document but the OCR engine is 75% sure that
the “o” is actually an “a”, the finished document would have the word “hause”. Turning this option on
tells the OCR engine to look at the other letters in the word, check to see which version of the word is in
the dictionary, and output the correct word “house” in the final document.
The OCR engine does not automatically correct misspelled words that were present in the original
document.
When dictionaries are selected, the terms in those dictionaries are used to check the spelling. If this is
option is not selected, User Dictionaries and Professional Dictionaries cannot be selected.
– User Dictionary —a user dictionary is your personal dictionary with words that you want the OCR engine
to reference for better accuracy when converting the document into editable text. For example, if you
scan documents with highly technical terms or acronyms not found in typical dictionaries, you can add
them to your personal dictionary. You can also add names that you expect to be in the documents too.
This way, as the OCR process recognizes each letter or symbol, there is a higher chance that the technical
term or name will be correctly spelled in the final document. You can create multiple user dictionaries.
See the section
Creating Your Own Dictionaries on page 95.
Click the menu arrow and select a user dictionary from the list.
If you select [none] as the user dictionary, the text will be validated using the terms in the dictionaries
for the selected languages, as well as any professional dictionaries if they are selected.
The label [current] is next to the currently-select user dictionary.
– Professional Dictionaries —these are legal and medical dictionaries containing highly specialized words
and phrases. The options are: Dutch Legal, Dutch Medical, English Financial, English Legal, English
Medical, French Legal, French Medical, German Legal, and German Medical. Select the appropriate
dictionary for the OCR engine to use to validate the scanned text.
– Reject Character —this is the character that the OCR process inserts for an unrecognizable text
character. For example, if the OCR process cannot recognize the J in REJECT, and ~ is the reject
character, the word would appear as RE~ECT in your document. The ~ is the default reject character.
Type the character you want to use in the Reject Character box. Try to choose a character that will not
appear in your documents.
– Missing Character —this is the character that the OCR process inserts for a missing text character. A
missing text character is one that the OCR process recognizes, but cannot represent because that
character is not available for the selected language. For example, if the document contains the text
symbol “Ç” but the OCR process cannot represent that character, then every place “Ç”
appears, the OCR
process substitutes the missing character symbol. The caret (^) is the default symbol for the missing
character.
Type the character you want to use in the Missing Character box. Try to choose a character that will not
appear in your documents.
– Recognition Quality —drag the slider to the left or right to set the degree of accuracy for the OCR
process. The higher the accuracy, the longer the OCR process requires to complete. For clean, highly-
legible documents, you can set the recognition quality to a lower level to produce results more quickly.