abbyy-software formreader ユーザーガイド
Identification of different forms processed in the same batch
There are certain things that must be considered during form creation:
•
whether the form is to be a multipage form
•
whether the form will be processed in the same batch with forms of a different type
In both cases, additional identification reference blocks are required. These elements allow the system to identify the
form type and select the proper template as well as to match it correctly (i.e. to locate fields location). The following
elements may be used:
1) Black squares
As mentioned previously, the optimal number of black squares on a form is five. Four of these squares should be
located in each corner, creating an imaginary rectangle, and the fifth on the side of this rectangle, its location
differing according to form page and form type (the fifth square being the feature that allows the application to
distinguish between different form pages and types). Note that fifth squares must differ in location by at least 10-15
mm if form identification is to be successful.
2) Barcodes
We recommend that EAN 13 barcodes 47- 50 mm in width (the distance between the barcode line furthest to the
right and the barcode line furthest to the left) and 12-14 mm in height (barcode digit heights are not included in
these measurements) be used, and that a distance of no less than 10 mm be allowed between each barcode and all
other form elements.
The barcode line direction should coincide with form page orientation, and the forms should be scanned in the same
direction as that of the barcode bars.
Barcodes located on different form pages as well as on forms of different types should always differ in value, and
only then can they be used as form identifiers.
3) Static text
The text should be clear and legible, and present no problems to an OCR system.. Static text identifiers should be
lines set in a plain monospaced font (without any stylization) and no less than 8 mm in size. We recommend a
distance of no less than 10 mm between each static text block and all other form elements. Static text lines located
on forms of different types should always differ in content, and only then they may be used as form identifiers.
Notes.
1. The number of reference points used on a form depends on the form design and the amount of free space
form type and select the proper template as well as to match it correctly (i.e. to locate fields location). The following
elements may be used:
1) Black squares
As mentioned previously, the optimal number of black squares on a form is five. Four of these squares should be
located in each corner, creating an imaginary rectangle, and the fifth on the side of this rectangle, its location
differing according to form page and form type (the fifth square being the feature that allows the application to
distinguish between different form pages and types). Note that fifth squares must differ in location by at least 10-15
mm if form identification is to be successful.
2) Barcodes
We recommend that EAN 13 barcodes 47- 50 mm in width (the distance between the barcode line furthest to the
right and the barcode line furthest to the left) and 12-14 mm in height (barcode digit heights are not included in
these measurements) be used, and that a distance of no less than 10 mm be allowed between each barcode and all
other form elements.
The barcode line direction should coincide with form page orientation, and the forms should be scanned in the same
direction as that of the barcode bars.
Barcodes located on different form pages as well as on forms of different types should always differ in value, and
only then can they be used as form identifiers.
3) Static text
The text should be clear and legible, and present no problems to an OCR system.. Static text identifiers should be
lines set in a plain monospaced font (without any stylization) and no less than 8 mm in size. We recommend a
distance of no less than 10 mm between each static text block and all other form elements. Static text lines located
on forms of different types should always differ in content, and only then they may be used as form identifiers.
Notes.
1. The number of reference points used on a form depends on the form design and the amount of free space
present on it. The presence of a large number of reference points on a form does not mean that all of them will
be used as such by the OCR system.
If the quantity of forms is to be very large (and especially if the form is to be printed professionally), we
recommend that the form be tested beforehand by printing out a copy on a printer, and creating an appropriate
template using the OCR system. Ensure that the reference points placed on the form result in both correct
template matching and deskewing (in the case of scanning defects), and correct form identification if several
types of form are present in the same batch.
If there is no possibility of testing a form before it is printed (samples of the other forms to be processed at the
same time, for example, may not be available), we recommend that two or more types of reference blocks be
used on the form concerned. As a result, a combination of reference points that provides the best form
identification can be used during template creation.
be used as such by the OCR system.
If the quantity of forms is to be very large (and especially if the form is to be printed professionally), we
recommend that the form be tested beforehand by printing out a copy on a printer, and creating an appropriate
template using the OCR system. Ensure that the reference points placed on the form result in both correct
template matching and deskewing (in the case of scanning defects), and correct form identification if several
types of form are present in the same batch.
If there is no possibility of testing a form before it is printed (samples of the other forms to be processed at the
same time, for example, may not be available), we recommend that two or more types of reference blocks be
used on the form concerned. As a result, a combination of reference points that provides the best form
identification can be used during template creation.
2. If a barcode or static text is to be used as a reference point, attention must be paid to their color. They must be
retained on form image after scanning, and their images must be of high quality, with no skew, garbage, or
glued bars/letters present. The recommended color for reference blocks is black.
glued bars/letters present. The recommended color for reference blocks is black.