Template creation advices

ABBYY FormReader 4.1

How to scan the template of a drop-out form?

When you use black&white forms there are no problems with getting the good image on the screen for the template creation and setup as all the fields, their names and other information is retained after scanning. But if you scan drop-out forms in the usual mode all additional information and field margins having the same color as the background disappear. To get the image with visible fields margins etc. you may mark the outlines of these by hand before scanning, or scan the forms in many shades of gray (if you do not scan the image from ABBYY FormReader and the image is saved into a file, choose the Gray TIFF Packbits format to save the file). The acquired image may be used to set up the template. To test and further input forms you must replace the gray scanning mode for the black&white (LineArt) mode.

In what cases it is not recommended to draw the reference blocks automatically?

1. In case there are black lines on the form image margins that lack on the source form. These lines will be automatically made reference blocks, but these lines appear on image only at certain scanning parameters and, so, may not be present on the entire set of forms. If they are made reference blocks the application will not be able to match the template for certain images.

Try to scan the form again, so that there are no lines, and replace the template image or draw the reference blocks manually.

2. The automatic drawing procedure always selects all the static text and all separators present on the image. But it is much more than needed for correct template matching. The excess number of reference blocks slows down the template matching and so slows down the whole form processing process. So, if there are many blocks with static text and/or separators, we recommend you to draw the blocks manually, and only those that are necessary.

3. There are cases when the automatic reference block drawing procedure draws every word of static text as a reference block. So if a static text contains more than one word, it is better to draw the respective reference block manually.

How to draw the reference blocks in such a way as to achieve the best template matching?

1. Only the form elements that are present on all forms irrespectively of the scanning mode and options may be made reference points.

2. If you process several types of forms in one batch (say, multipage forms) you should position the reference blocks on different templates in such a way that different forms have different locations of reference points, or you should draw some blocks of "static text" or "barcode" type and specify for them different values for different forms.

What the size of a text field for recognition must be?

When you draw a text field or checkmark, we recommend you to draw a little bit larger block (by some 1-2mm) than the one on the form image, as the application will only recognize the text that is located near margins or outreaches them by a little in case the block encompasses it entirely. So the size of the template block must be a little bit greater than the size of the source form field.

How to change the block type and its properties quickly?

If you do not close the Properties dialog when working in Template Editor, you can just click a block to access its properties - they will be immediately displayed in the dialog on the Block tab.

How to draw a "Checkmark" block correctly?

A checkmark block always have a rectangle form in the template, irrespectively of its actual form on the image: circle, square, oval etc. When you draw a Checkmark block, make sure that the checkmark image together with its margins were enclosed by the respective template block entirely.

How to create a block of «Checkmark» type for a checkmark in a form of a text to be enclosed in a circle?

It is not good to have such checkmarks on machine-readable forms. But if your forms already contains such checkmarks and you cannot correct this, you may draw a "checkmark" block near the text (to the left or lower than it), so as to have the enclosing circle cross the block. Yet the block must be separated from the text by a certain distance, like this:

For example:

How can I set up the batch options correctly if I scan raster forms or forms with dirty or gray background?

If you set up the batch to process raster forms, do not forget to check the Clean images when opening option. In this case the batch images will have only the data written in the text fields and reference points.

Also, this option is useful for forms for which you cannot set up the scanning options in such a way so as to acquire a garbage-free image.

Note. If the source form text is very light or very thin font was used on the form, and also if the form has been filled in in a very light-colored pen, the use of the Clean image option may result in disappearance of fullstops, commas and small or thin parts of letters. This may result in a low recognition quality.

If the form background is not smooth (somewhere "dirty", somewhere light), you should clean various form blocks and not the entire form. The block cleaning option is set up in the block parameters during the template editing (Template editor, Properties>Block>Advanced options>Clean block).

If you cannot find answers to your questions on this website, please feel free to contact our technical support service. Click on the following link to open a list of countries where technical support is available in local languages: Support Contacts

<< Back