If your solution is licensed for intelligent character recognition (ICR), you can use Advanced Capture to recognize and capture data from hand-written text.
It is difficult to predict the accuracy of an ICR process due to the variability in the quality of the hand-written text. The following tips will help to improve the accuracy of ICR processing.
-
The hand-written characters should be clearly separate from one another (i.e., they should not touch or intersect).
-
The hand-written text should be unconstrained. If the characters are being entered into constraint boxes, do one of the following:
-
If the documents containing the constraint boxes can be modified before the Advanced Capture process is run, ensure that the constraint boxes are printed in red. Then, using Document Imaging's Image Processing options, configure a red Color Dropout process to remove the constraint boxes from the documents when scanning.
-
If the documents containing the constraint boxes cannot be modified before the Advanced Capture process is run, configure your OCR process format to automatically drop the constraint boxes by selecting Remove vertical/horizontal lines from zone before processing in the Data Field Zone dialog box.
Constrained Text:
Unconstrained Text:
-
-
The ICR numeric text processor (Handwriting/ICR - numerals/grouping punctuation (North American style), Handwriting/ICR - numerals/grouping punctuation (European style)) is significantly more accurate than the ICR alphanumeric text processor (Handwriting/ICR - alphanumeric/punctuation) because the character set is much smaller.
If you are attempting to process numeric data only from a Data Field Zone, you will have more accurate results if you use the numeric text processor.