User Guide
Automated forms processing: step by step
Designing a form
First of all you have to design a form. You need a form that is both
easy to fill in and to process. The design is crucial because any mistakes
made at this stage may drastically reduce the speed of processing. Be
sure to follow the recommendations of the supplier of your data cap
ture application.
To create a form, you first need to think out its logical structure,
then design it, and, finally, draw your form. A detailed treatment of
each stage follows.
Determining the form's logic
Forms with carefully thoughtout logical structures are easier
to fill in and process. You need to decide what data you will need
to gather and draw up a list of required data fields. Be sure to dis
cuss these fields with your clients.
Next, you have to determine such parameters as the size of the
form and the number of pages in the form. If you decide to change
these parameters later, you may need to make considerable changes
in the setup of your data capture program. This is why we recom
mend drawing all your sketches on sheets of paper the size of your
future form so that you have enough space for all the elements you
wish to place on the form.
ID fields (identifiers).
If you need a form which consists of
several pages, be sure to introduce elements which will help you
avoid confusion. Usually each page is provided with an ID field
which is the same for all the pages of the same kind. The nature of
the ID field depends on the nature of the form. This can be an SSN,
a customer's ID, the code of the project, etc.
Simple and complex fields.
Try making your fields as simple
as possible. This will make for fewer errors when filling in and pro
cessing the form. The more predictable the words or numbers
entered into particular fields, the higher the recognition rate. It is
best to split such fields as Name, Date, Telephone (area code + city
code), Address (country+city+street)into several subfields.
Free space is a rare commodity on any form, therefore if you
know the maximum length of a field do not make it larger than
necessary. This will prevent the person filling in the form from
entering redundant information and will make the whole comple
tion process more selfevident. Examples of data fields with a
known number of character spaces: SSN, postal code, abbreviations
for US states, local telephone numbers, standard names of curren
cies.
Field length.
The length of words in such fields as Street,
Second Name or Email is difficult to predict, therefore you need to
provide some extra character spaces "just in case". If you think one
line may prove insufficient, allocate two or more lines to such fields.
FormReader can logically merge such lines into one field without
diminishing recognition quality
.
Separators.
The form must encourage people who fill it in to
enter only meaningful information into its fields. For example, it
would be wise to design a Date field in such a way that users do not
have to enter separators (e.g. slashes, hyphens or dots) themselves,
because they will be printed on their forms. This will greatly increase
recognition accuracy. Similar examples: preprinted hyphens in SSN
and ISO fields, the first three digits of the current year, etc.
Check boxes.
If possible answers are known in advance, it is
best to use check boxes instead of text fields, as OMR algorithms
are much more reliable than ICR. For example, do not ask users to
fill in their marital status in text fields by writing such words as
"married", "single", "divorced" or "widower". Instead, print the possi
ble answers on the form and ask users to tick the appropriate box.
Captions and photos.
If your form will include such fields as
Signature, Seal, Photo or Fingerprint, be sure to provide enough
space for these fields. This will reduce the number of corrections
and increase recognition quality. Don't forget that affixing a stamp
or putting a seal on a form may result in blots on the reverse side
which may impede the recognition of the text there.
Selecting form type and design
One of the major recognition tasks is to separate the contents
of the fields from the field boundaries. The success of this task
largely depends on selecting the right type of form. Remember that
colour dropout forms provide the best results. Users will enter
information into white rectangles and the scanner will later remove
the background. The general rule of thumb: use grey forms when
ever you cannot print colour dropout forms.
When designing your forms, pay particular attention to refer
ence points and ID fields this will help you get the most out of
automated forms processing.
What is a reference point?
FormReader uses reference points
to match forms with their templates. Reference points are also used
to correct linear distortions introduced by scanning and to detect
the location of the fields on the form. Sometimes reference points
are referred to as anchors. Examples of reference points: black
squares, corners, crosses, captions that do not disappear during
scanning, vertical and horizontal lines. We recommend placing three
or four reference points in the corners of the page. This will enable
FormReader to match forms with their templates and to process
similar forms printed on different printers or sent in by fax.
What is an identifier?
Identifiers are form elements that do not
disappear during scanning and that are used to match a form with its
template. If multiplepage forms are processed within one batch, you
need to provide a unique element on each page which will be used
to identify pages as belonging to a particular form. We recommend
using bar codes, form titles or additional black squares as form iden
tifiers.
Drawing a form
Once the logical structure of the form has been arrived at, you
need to draw your form. What drawing tool should you use?
Currently there are several tools available on the market.
If your designer is familiar with CorelDRAW or Adobe
Illustrator, they may draw the form in one of these applications.
These are good professional design tools but they have their draw
backs both programs are a bit too "heavy" and expensive. They are
too difficult for a novice to learn, and learning all their features will
take considerable time.
Microsoft Visio is more common and less difficult to use. Even
though it is mainly intended for drawing charts and graphics, it can
also be used to draw quite attractive forms. The easiest way to
design a form in MS Visio is to use its socalled template galleries.
You can obtain a template gallery containing various form elements
form ABBYY. MS Visio can be used to create professionallylooking
colour dropout forms which can then be printed on a laser printer.
As a last resort, you can create forms in Microsoft Word. Since
MS Word was not originally intended as a tool for designing forms,
drawing a form in this text editor may prove a real challenge.